Patient post-diagnosis intelligent follow-up visit method and device, storage medium and program product

Abstract

The embodiment of the invention discloses an intelligent follow-up visit method and device for a patient after diagnosis, a storage medium and a program product. The method comprises the following steps: creating an employee list and a patient list; determining push time based on the patient list; when the push time arrives, selecting an optimal medical worker based on the staff list and the patient list; a follow-up visit task is pushed to the terminal device of the optimal medical worker; acquiring medical data of the patient in response to a follow-up visit request of a medical worker for the patient; performing health detection on the patient based on the medical data; if abnormity occurs, early warning information is generated, treatment guidance is conducted on the patient based on the early warning information, and the follow-up visit task is pushed again. According to the method, automatic follow-up task pushing, remote follow-up and monitoring can be achieved, labor cost is reduced, follow-up work efficiency, quality and comprehensiveness are improved, more convenient medical services are provided for patients, and medical resources can be effectively utilized.

Description

Technical Field

[0001] The present disclosure relates to the field of medical health technology, and in particular to a method, device, storage medium, and program product for intelligent follow-up of patients after diagnosis. Background Art

[0002] In the medical field, follow-up is an important part of ensuring that patients receive continuous care and monitoring. Traditional follow-up methods mainly rely on manual records and telephone communication, which are not only inefficient but also prone to errors, making it difficult to ensure the quality of follow-up. In addition, due to limited medical resources and busy medical staff, it is difficult to ensure the timeliness and comprehensiveness of follow-up, resulting in low efficiency of follow-up work, high loss rate, and loss of a large amount of valuable medical data. Summary of the invention

[0003] In view of this, the embodiments of the present disclosure provide a method, device, storage medium, and program product for intelligent follow-up of patients after diagnosis, which can realize automatic push of follow-up tasks, remote follow-up and monitoring, improve the efficiency and comprehensiveness of follow-up work, and provide patients with more convenient medical services.

[0004] In a first aspect, the present disclosure provides a method for intelligent follow-up of patients after diagnosis, which adopts the following technical solutions:

[0005] Create staff and patient lists;

[0006] Determining a push time based on the patient list;

[0007] When the push time arrives, the best medical staff is selected based on the employee list and the patient list;

[0008] Pushing the follow-up task to the terminal device of the optimal medical staff;

[0009] Responding to a medical staff follow-up request for a patient, obtaining the patient's medical data;

[0010] Conduct health checks on patients based on medical data;

[0011] If an abnormality occurs, an early warning message is generated, treatment guidance is given to the patient based on the early warning message, and a follow-up task is re-sent.

[0012] Optionally, the calculation formula for the push time is:

[0013] T 推送 =T i就诊 +X*T 基础 -T 浮动 ;

[0014] Where, T 推送is the push time; i is the patient's visit number; T i就诊 is the patient's consultation time; the patient list includes the patient's consultation number and consultation time; X is the current follow-up number; T 基础 is the preset base date; T 浮动 The default floating date.

[0015] Optionally, the calculation formula for the push time is:

[0016] T 推送 =T i就诊 +X*(T 基础 -T j )-T 浮动 ;

[0017] Where, T 推送 is the push time; i is the patient's visit number; T i就诊 is the patient's visit time; X is the current follow-up number; T 基础 is the preset base date; j is the severity level; T j T is the advancement time of the severity level; 浮动 It is a preset floating date; the patient list includes the patient's visit number, visit time and severity level.

[0018] Optionally, based on the employee list and the patient list, a comprehensive matching score between each medical staff and the patient is obtained, and the medical staff with the highest comprehensive matching score is selected as the best medical staff;

[0019] The expression of the comprehensive matching score is as follows:

[0020] FH=w S ·S+w W ·W+w H ·H+w P ·P+w F ·F;

[0021] Wherein, FH is the comprehensive matching score; w S 、w W 、w H 、w P 、w F is the preset weight; S is the expertise matching score; W is the workload score; H is the historical interaction record score; P is the patient demand score; F is the familiarity score;

[0022]

[0023] Where n is the total number of disease types, 1 <k≤n;p kis the kth element in the medical staff expertise vector, indicating whether the medical staff has expertise in treating the kth disease. If so, p k =1, if not, then p k =0;d k is the kth element in the patient's disease type vector, indicating whether the patient suffers from the kth disease. If so, d k =1, if not suffering from, then d k =0;

[0024]

[0025] Among them, Z is the current follow-up workload; Zm is the historical maximum follow-up workload; a, b, and c are all preset adjustment parameters;

[0026]

[0027] Among them, λ is the initial interaction score; e is a natural constant; t is the number of days experienced by the medical staff and the patient from the last actual follow-up time to the current time; μ is the decay parameter;

[0028]

[0029] Among them, β0 is the basic demand score; β1 is the degree of influence of the patient's age on the demand; L is the patient's age; β2 is the degree of influence of the patient's disease type on the demand;

[0030]

[0031] Where V is the current total number of statistical cycles; x v is the frequency of follow-up visits of medical staff to patients during the vth statistical period; is the average follow-up frequency in all statistical periods; v is the patient satisfaction in the vth statistical period; It is the average of the patients' satisfaction during all statistical periods.

[0032] Optionally, in response to a medical staff's request for a follow-up visit to the patient, obtaining the patient's medical data includes:

[0033] In response to a follow-up request from a medical staff to a patient, providing the medical staff with a follow-up questionnaire template;

[0034] Responding to a submission request from a medical staff member, obtaining a completed follow-up questionnaire template;

[0035] Extracting information from the completed follow-up questionnaire template to obtain the patient's current condition data;

[0036] Access to clinical data of patients across several healthcare systems;

[0037] Dividing the current condition data and the clinical data according to data categories to obtain a number of text data, numerical data and medical images;

[0038] The plurality of text data, numerical data and medical images constitute the medical data.

[0039] Optionally, the health check of the patient based on the medical data includes:

[0040] Performing local feature extraction on the plurality of text data, numerical data and medical images respectively to obtain a plurality of text features, numerical features and image features;

[0041] Compressing the plurality of text features, numerical features and image features respectively, and transmitting the compressed features to a computing center;

[0042] The compressed features are decompressed by the computing center, and health checks are performed on the patient based on the decompressed text features, numerical features, and image features.

[0043] Optionally, if the medical image is two-dimensional, converting the medical image into three-dimensional;

[0044] Use 3D convolutional neural network to extract features from three-dimensional medical images and obtain image features;

[0045] The 3D convolutional neural network includes an input layer and a convolution layer;

[0046] The input layer is used to receive a three-dimensional medical image;

[0047] The convolutional layer is used to extract features from the received medical image;

[0048] YY+242248P

[0049] The convolutional layer includes three 3D convolutional layers;

[0050] The first two 3D convolutional layers are connected to the 3D maximum pooling layer respectively;

[0051] The last 3D convolution layer is connected to the global average pooling layer and two fully connected layers;

[0052] ReLU activation function is applied to the output of each 3D convolutional layer and the output of the two fully connected layers.

[0053] Optionally, the performing health check on the patient based on the decompressed text features, numerical features and image features includes:

[0054] If the number of the image features is equal to 1, constructing a comprehensive three-dimensional image based on the text features, the numerical features and the image features;

[0055] Based on the comprehensive three-dimensional image, performing health examination on the patient;

[0056] If the number of the image features is not equal to 1, a fusion level is assigned to each image feature, and the fusion level is 0;

[0057] Determining whether the number of the image features is odd;

[0058] If it is an odd number, the fusion level of any image feature is increased by 1, and image features with the same fusion level are fused in pairs to obtain a number of new image features, and it is re-determined whether the number of the image features is equal to 1;

[0059] If it is not an odd number, image features of the same fusion level are fused in pairs to obtain a number of new image features, and it is re-determined whether the number of the image features is equal to 1.

[0060] Optionally, if no abnormality is found in the health check, it is determined that the follow-up task is completed;

[0061] determining whether the patient has recovered;

[0062] If not recovered, the follow-up task will be pushed again;

[0063] If the patient recovers, the follow-up for the patient is considered to be over.

[0064] In a second aspect, the present disclosure also provides a patient post-diagnosis intelligent follow-up system, which adopts the following technical solutions:

[0065] Create modules for creating staff lists and patient lists;

[0066] A determination module, used for determining a push time based on the patient list;

[0067] A selection module, configured to select the best medical staff based on the staff list and the patient list when the push time is reached;

[0068] A push module, used for pushing the follow-up task to the terminal device of the optimal medical staff;

[0069] An acquisition module, used to obtain the patient's medical data in response to a follow-up request from a medical staff member;

[0070] The detection module is used to perform health detection on patients based on medical data;

[0071] The generation module is used to generate warning information if an abnormality occurs, provide treatment guidance to the patient based on the warning information, and re-send the follow-up task.

[0072] In a third aspect, the embodiments of the present disclosure further provide a computer device, which adopts the following technical solution:

[0073] The computer device comprises:

[0074] at least one processor; and,

[0075] a memory communicatively connected to the at least one processor; wherein,

[0076] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the above-mentioned intelligent follow-up methods for patients after diagnosis.

[0077] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute any of the above-mentioned intelligent follow-up methods for patients after diagnosis.

[0078] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, including a computer program / instruction, which implements the steps of any of the above methods when executed by a processor.

[0079] The intelligent follow-up method for patients after diagnosis provided by the disclosed embodiment reduces the time and effort of manually arranging follow-up work by automatically pushing follow-up tasks to medical staff, thereby improving the efficiency of follow-up work, and ensuring that each patient can receive targeted follow-up services by assigning personalized tasks according to the employee list and patient list. By collecting and analyzing the patient's medical data, more accurate health test results can be provided, thereby supporting medical staff to make more data-based clinical decisions. When abnormalities occur in the patient's health test, early warning information can be generated in time, prompting medical staff to take treatment measures quickly, thereby improving the response speed of the patient's health risks, and re-pushing follow-up tasks ensures that medical resources can be more effectively allocated to patients who need attention, thereby improving resource utilization efficiency. This timely follow-up and early warning allows patients to feel more intimate and professional medical services, thereby improving patients' satisfaction with medical services, and professional treatment guidance can help medical staff treat patients more effectively, thereby improving treatment effects and patient recovery rates. This method not only focuses on short-term treatment effects, but also promotes patients' long-term health management through long-term follow-up and health testing. In the context of the growing development of telemedicine and digital medicine, this method supports remote follow-up and monitoring, providing patients with more convenient medical services.

[0080] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following preferred embodiments are specifically cited and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0082] Figure 1 A schematic diagram of the process of the intelligent follow-up method for patients after diagnosis provided by the embodiment of the present disclosure;

[0083] Figure 2 A schematic diagram of a follow-up system interface provided by an embodiment of the present disclosure;

[0084] Figure 3 Another schematic diagram of the follow-up system interface provided by the embodiment of the present disclosure;

[0085] Figure 4 A flowchart of a medical data acquisition method provided in an embodiment of the present disclosure;

[0086] Figure 5 A flowchart of a health detection method provided by an embodiment of the present disclosure;

[0087] Figure 6 A schematic diagram of a process of constructing a three-dimensional image provided by an embodiment of the present disclosure;

[0088] Figure 7 A principle block diagram of a patient post-diagnosis intelligent follow-up system provided by an embodiment of the present disclosure;

[0089] Figure 8 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0090] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0091] It should be clear that the following embodiments of the present disclosure are described by specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0092] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structures and / or functions described herein are merely illustrative. Based on this disclosure, those skilled in the art should understand that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of the aspects described herein may be used to implement the device and / or practice the method. In addition, one or more of the aspects described herein may be used.

[0093] Other structure and / or functionality other than may be used to implement this apparatus and / or practice this method.

[0094] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0095] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0096] Reference Figure 1 The present disclosure provides a method for intelligent follow-up of patients after diagnosis, comprising the following steps:

[0097] S1: Create staff and patient lists;

[0098] S2: Determine the push time based on the patient list;

[0099] S3: When the push time arrives, the best medical staff is selected based on the staff list and patient list;

[0100] S4: Push the follow-up task to the terminal device of the best medical staff;

[0101] S5: Responding to a follow-up request from a medical staff member to a patient, obtaining medical data of the patient;

[0102] S6: Perform health checks on patients based on medical data; if an abnormality is found, execute S7; if no abnormality is found, execute S8:

[0103] S7: Generate early warning information, provide treatment guidance to patients based on the early warning information, and re-send follow-up tasks;

[0104] S8: Determine that the follow-up task is completed.

[0105] The disclosed patient post-diagnosis intelligent follow-up method reduces the time and effort of manually arranging follow-up work by automatically pushing follow-up tasks to medical staff, thereby improving the efficiency of follow-up work, and ensuring that each patient can receive targeted follow-up services by assigning personalized tasks based on employee lists and patient lists. By collecting and analyzing the patient's medical data, more accurate health test results can be provided, thereby supporting medical staff to make more data-based clinical decisions. When abnormalities occur in the patient's health test, early warning information can be generated in a timely manner, prompting medical staff to quickly take treatment measures, thereby improving the response speed of the patient's health risks, and re-pushing follow-up tasks ensures that medical resources can be more effectively allocated to patients who need attention, thereby improving resource utilization efficiency. This timely follow-up and early warning allows patients to feel more intimate and professional medical services, thereby improving patient satisfaction with medical services, and professional treatment guidance can help medical staff treat patients more effectively, thereby improving treatment effects and patient recovery rates. This method not only focuses on short-term treatment effects, but also promotes patients' long-term health management through long-term follow-up and health testing. In the context of the growing development of telemedicine and digital medicine, this method supports remote follow-up and monitoring, providing patients with more convenient medical services.

[0106] In summary, this intelligent follow-up method can effectively reduce unnecessary outpatient visits and hospitalizations, reduce medical costs, and improve the utilization efficiency of medical resources. Through real-time monitoring and data analysis, it can promptly detect patients' health problems and take intervention measures, effectively manage chronic diseases and improve treatment effects. Through communication through multiple channels and regular follow-up, it can achieve more convenient and personalized medical services, enhance communication and understanding between doctors and patients, and improve treatment compliance and treatment effects. It is an important part of the modern medical system.

[0107] In S1, an employee list and a patient list are created in the follow-up system. The employee list includes the employee information of multiple medical staff, and the employee information includes at least one of the medical staff's name, work number, department, specialty, current follow-up workload, historical maximum follow-up workload, and actual follow-up time (including the actual time of all follow-ups). The patient list includes the basic information of multiple patients, and the basic information includes at least one of the patient's name, mobile phone number, ID number, age, consultation time, consultation number, disease type, and severity level.

[0108] The basic information of patients comes from several medical systems. The basic information of patients in different medical systems is transmitted to the follow-up system. The follow-up system integrates this basic information and stores it in the patient list. As the number of patients increases or decreases, the basic information of patients in the medical system will be updated in real time, and accordingly, the patient list will also be updated in real time. In the process of basic information transmission, encryption technology and security measures are used to ensure that patient data is securely protected and in compliance with relevant laws, regulations and privacy policies.

[0109] In S2, the first feasible implementation plan is: the medical institution administrator presets a follow-up plan, which includes a basic date and a floating date. The push time is calculated according to the basic date, the floating date and the patient's consultation time. The calculation formula of the push time is as follows:

[0110] T 推送 =T i就诊 +X*T 基础 -T 浮动 ;

[0111] Where, T 推送 is the push time; i is the patient's visit number; T i就诊 is the patient's visit time; X is the current follow-up number; T 基础 is the base date; T 浮动 A floating date.

[0112] For example, if the base date is 3 months, the floating date is 7 days, and the patient's visit date is January 1, 2024, after calculation, the push date of the first follow-up is March 26, 2024, and the task period is set to 14 days. Its essence is: set a 3-month follow-up cycle, and on this basis set a floating time of ±7 days for medical staff to complete the follow-up task. This method facilitates medical institution administrators to control the layout of follow-up tasks through the follow-up plan, while ensuring that medical staff can give full play to their professional advantages and provide high-quality medical services.

[0113] The second possible implementation plan is: the medical institution administrator presets a follow-up plan, which includes a basic date and a floating date. The push time is calculated according to the basic date, floating date, patient's visit time and severity level. The calculation formula for the push time is as follows:

[0114] T 推送 =T i就诊 +X*(T 基础 -T j )-T 浮动 ;

[0115] Where j is the severity level; T j The advancement time of severity level. Different severity levels have different advancement time. For example, the severity level is divided into 3 levels. The advancement time of level 1 is 0, the advancement time of level 2 is 1 month, and the advancement time of level 1 is 2 months.

[0116] In S3, a feasible solution is: when the current time reaches the push time, based on the patient's disease type and the expertise of all medical staff in the staff list, select a number of suitable medical staff, that is, select the medical staff suitable for the patient's disease type according to their expertise; if the number of suitable medical staff is greater than 1, select the medical staff with the smallest current follow-up workload as the best medical staff; if the number of suitable medical staff is equal to 1, the suitable medical staff is also determined to be the best medical staff. Push the follow-up task to the terminal device of the best medical staff, add 1 to the current follow-up number, set a push reminder, and add a task deadline for the follow-up task. The task deadline can be initially set in the follow-up plan.

[0117] In S3, another feasible solution is: based on the employee list and patient list, obtain the comprehensive matching score between each medical staff and the patient, and select the medical staff with the highest comprehensive matching score as the best medical staff.

[0118] The expression of the comprehensive matching score is as follows:

[0119] FH=w S ·S+w W ·W+w H ·H+w P ·P+w F ·F;

[0120] Among them, FH is the comprehensive matching score; w S 、w W 、w H 、w P 、w F is the preset weight, which can be determined by expert scoring, analytic hierarchy process (AHP) or other methods. S 、w W 、wH 、w P 、w F ∈(0, 1) and their sum is 1; S is the expertise matching score, which is used to measure the matching degree between the expertise of medical staff and the disease needs of patients; W is the workload score, which is an indicator to measure the current workload of medical staff; H is the historical interaction record score, which is used to quantify the historical interaction frequency between medical staff and patients (also known as historical follow-up frequency); P is the patient demand score, which is used to quantify the patient's demand for medical resources; F is the familiarity score, which is an indicator to measure the familiarity between medical staff and patients.

[0121]

[0122] Among them, k is an index representing the disease type; n is the total number of disease types, 1 <k≤n;p k is the kth element in the medical staff expertise vector, indicating whether the medical staff has expertise in treating the kth disease. If so, p k =1, if not, then p k = 0, for example, a medical professional specializes in cardiology and the patient has heart disease, p k is 1, otherwise it is 0; d k is the kth element in the patient's disease type vector, indicating whether the patient suffers from the kth disease. If so, d k =1, if not suffering from, then d k = 0, for example, the patient has heart disease, d k is 1 if the value is true, otherwise it is 0.

[0123]

[0124] Wherein, Z is the current follow-up workload; Zm is the historical maximum follow-up workload; a, b, and c are all preset adjustment parameters. Wherein, a is used to control the overall size of the workload score, which is determined by historical data analysis to ensure that the workload score is within a reasonable range. The initial a is 1. As the historical data increases, the value of a is adjusted. b is used to control the degree of influence of the current follow-up workload on the workload score. It determines whether the relationship between the current follow-up workload and the workload score is linear or nonlinear. In this disclosure, b takes a value of 2. c is used to ensure that when the current follow-up workload is 0, the workload score is not negative. In this disclosure, c takes a value of 5.

[0125]

[0126] Among them, λ is the initial interaction score; e is a natural constant; t is the number of days experienced by the medical staff and the patient from the last actual follow-up time to the current time (that is, the time since the last interaction between the medical staff and the patient); μ is a decay parameter used to control the decay speed of the interaction record, which determines the rate of reduction of the impact of historical interaction records on the current historical interaction record score. In this disclosure, μ is taken as 30.

[0127]

[0128] Among them, β0 is the basic demand score, which is a constant term, representing the basic demand level of the patient without any other factors (that is, when age and disease type are not considered); β1 is the degree of influence of the patient's age on the demand, which is a coefficient used to measure the influence of age on the patient's demand; L is the patient's age; β2 is the degree of influence of the patient's disease type on the demand, which is a coefficient used to measure the influence of the disease type on the patient's demand. β0, β1, and β2 are determined by data fitting. In this paper, β0 = 0, β1 = 0.05, and β2 = 1.

[0129]

[0130] Where V is the current total number of statistical cycles; x v is the frequency of follow-up visits of medical staff to patients during the vth statistical period; is the average follow-up frequency in all statistical periods; v is the patient satisfaction during the vth statistical period, which is specifically the average of the patient's satisfaction with all follow-up visits to medical staff during the statistical period; It is the average of the patients' satisfaction during all statistical periods.

[0131] When selecting the best medical staff, the above is quantified through the patient's historical follow-up records, the frequency of interaction between medical staff and patients, and the degree of medical staff's understanding of the patient's medical history. For example, record the number of follow-up visits and feedback from the medical staff with the patient in the past, and establish a scoring system to evaluate the medical staff's familiarity with the patient. According to the patient's disease type, age, underlying disease, and living habits, a comprehensive analysis is conducted. When selecting medical staff, those with experience in related fields are given priority. If necessary, the main and deputy medical staff can be configured. The main medical staff is responsible for the main follow-up tasks, while the deputy medical staff provides support to ensure that the professional knowledge reserves of the follow-up can meet the actual needs of the patients. This configuration can improve the comprehensiveness and professionalism of the follow-up. Through these measures, the personalized needs of patients can be better met, the quality and efficiency of follow-up can be improved, and the satisfaction of patients can also be improved.

[0132] Further, refer to Figure 2 The schematic diagram of the follow-up system interface is displayed. According to the completion progress of the follow-up tasks, the follow-up tasks are classified into overdue, in progress, and about to be overdue. The follow-up tasks in different statuses are displayed through an intuitive interface, and reminders are set for follow-up tasks that are about to be overdue and have been overdue, so as to urge relevant medical staff to complete the follow-up tasks as soon as possible.

[0133] Reference Figure 3 Another schematic diagram of the follow-up system interface is shown. The follow-up system is equipped with a screening function. Medical staff can arbitrarily set the patient's gender, start time, end time, status, etc. of the follow-up task to screen out the required follow-up tasks.

[0134] In S5, refer to Figure 4 The flowchart of the medical data acquisition method shown, "obtaining the patient's medical data in response to a follow-up request from a medical staff member for the patient" includes the following steps:

[0135] S51: In response to a follow-up request from a medical staff to a patient, providing the medical staff with a follow-up questionnaire template;

[0136] S52: Responding to the submission request of the medical staff, obtaining a completed follow-up questionnaire template;

[0137] S53: extracting information from the completed follow-up questionnaire template to obtain the patient's current condition data;

[0138] S54: Obtain clinical data of patients in several medical systems;

[0139] S55: dividing the current condition data and clinical data according to data categories to obtain a number of text data, numerical data and medical images;

[0140] S56: Several text data, numerical data and medical images constitute medical data.

[0141] In S51-S53, refer to Figure 2, medical staff select patients to be followed up through the follow-up system and check the patient's treatment information. When the medical staff selects a follow-up task for any patient through a mouse or other device, it triggers the medical staff's follow-up request for the patient. In response to the request, the follow-up questionnaire template is displayed to the medical staff's terminal device. The follow-up questionnaire template can be set uniformly, or multiple types of follow-up questionnaire templates can be set in advance. According to the patient's disease type, the best one is selected from multiple follow-up questionnaire templates for display to the medical staff. The follow-up questionnaire template includes symptom assessment, medication status, etc., for medical staff to refer to during follow-up. The follow-up system supports multiple remote follow-up methods such as telephone and video. When communicating with patients, medical staff fill in the follow-up questionnaire template through communication with patients to obtain the patient's recent condition. When the follow-up questionnaire template is completed, click the "Submit" button, and the follow-up system obtains the completed follow-up questionnaire template, YY+242248P

[0142] Information is extracted from the template to obtain the patient's current condition data.

[0143] In S54-S56, the medical system will archive the patient's clinical data in chronological order and send the patient's medical number to several medical systems used. The medical system will filter out the patient's latest clinical data by the medical number. Note that these clinical data are required to be the patient's latest data. For example, if the patient has taken two brain CT images, the medical system will automatically identify the latter CT image as the latest data. Moreover, these clinical data will not be directly transmitted to the follow-up system, but will be used for local calculations later to prevent the leakage of the patient's privacy. In terms of medical diagnosis, data types are divided into three categories: text, numerical and image. Clinical data and current condition data can be divided into three types: text data, numerical data and medical images according to different data types. Medical data includes these text data, numerical data and medical images.

[0144] In S6, refer to Figure 5 The flowchart of the health detection method shown, "Health detection of patients based on medical data", includes the following steps:

[0145] S61: extracting local features from a plurality of text data, numerical data and medical images respectively to obtain a plurality of text features, numerical features and image features;

[0146] S62: compressing a plurality of text features, numerical features and image features respectively, and transmitting the compressed features to a computing center;

[0147] S63: The compressed features are decompressed by the computing center, and health checks are performed on the patient based on the decompressed text features, numerical features, and image features.

[0148] In S61, natural language processing technology (such as word embedding or sentence embedding) is used locally to convert text data into numerical vectors. This technology can capture semantic information in the text and convert it into a numerical form that can be processed by the machine learning model. The numerical vector is mapped to a 3D space (for ease of distinction, the 3D space here is called a text 3D space) through the TF-IDF weight method for easy analysis and processing. A preset machine learning model (such as a random forest, a gradient boosting tree, etc.) is used to extract features in the text 3D space. The preset machine learning model can identify and learn patterns and relationships in text data, thereby extracting useful features, and all features extracted from a text data are recorded as text features. For each text data, the method of extracting features locally can reduce the risk of patient privacy leakage. In addition, if there are multiple text data locally, multiple text features will be obtained.

[0149] Statistical methods (such as principal component analysis, t-distribution random neighbor embedding or linear discriminant analysis) are used locally to map multidimensional numerical data into 3D space (for ease of distinction, the 3D space here is called numerical 3D space) to retain the main variability of the numerical data. A preset machine learning model (such as random forest, gradient boosting tree, etc.) is used to extract features in the numerical 3D space. The model can identify key patterns and relationships in the numerical data and extract features that reflect the main variability of the numerical data. All features extracted from one numerical data are recorded as numerical features. If there are multiple numerical data locally, multiple numerical features will be obtained.

[0150] Medical images are preprocessed locally. The preprocessing includes image resizing and image enhancement. Image resizing can ensure that all medical images have the same size, and image enhancement can improve the quality of medical images.

[0151] Optionally, the image enhancement step can use techniques such as local binary pattern (LBP), Gabor filter or wavelet transform to extract the texture and multi-scale spatial energy distribution characteristics of the image, which help to enhance different features of the image and provide richer information for subsequent feature extraction.

[0152] If the medical image is three-dimensional, a 3D convolutional neural network (3D CNN for short) is used to extract features from the medical image. All features extracted from a medical image are recorded as image features. If there are multiple medical images locally, multiple image features will be obtained. If the medical image is two-dimensional, the medical image is mapped to a 3D space (for ease of distinction, the 3D space here is referred to as the image 3D space), the medical image is converted from two-dimensional to three-dimensional, and a 3D CNN is used to extract features in the image 3D space (i.e., feature extraction is performed on the medical image converted to three-dimensional) to obtain image features.

[0153] Among them, 3D CNN includes input layer and convolution layer. The input layer is the entrance of the network and accepts three-dimensional medical images with a shape of (64,64,64,1), that is, each input sample is a single-channel volume data of 64*64*64 pixels. The convolution layer extracts features from the medical images received by the input layer. The convolution layer consists of three 3D convolution layers. Each 3D convolution layer uses a different number and size of filters. The first 3D convolution layer contains 32 3x3x3 filters to capture basic features in medical images, such as edges and corners. The second 3D convolution layer contains 64 3x3x3 filters to extract more complex features. The third 3D convolution layer contains 128 3x3x3 filters to further extract higher-level features. Each 3D convolution layer is followed by a ReLU (Rectified Linear Unit) activation function to introduce nonlinearity and enhance the network's expressiveness. A 3D maximum pooling layer is also used after the first two 3D convolutional layers, with a pooling window size of 2x2x2, to reduce the spatial dimension of the features, reduce the amount of computation, and extract significant features. A global average pooling layer and two fully connected layers are also used after the last 3D convolutional layer. The global average pooling layer is used to compress the spatial dimensions (width, height, depth) of each feature map into a single value, that is, the average value of each feature map, which helps to reduce the number of parameters of the fully connected layer. The first fully connected layer has 256 neurons, and the second fully connected layer has 128 neurons. These two fully connected layers also use the ReLU activation function to learn nonlinear combinations between features.

[0154] The input medical image first passes through a series of 3D convolutional layers, each followed by a ReLU activation function to introduce nonlinearity, allowing the network to learn and represent more complex features. The 3D maximum pooling layer is used to reduce the spatial dimensions of the features (i.e., width, height, and depth) while retaining the most important information, which helps reduce the amount of computation and improve the abstraction level of the features. After multiple 3D convolutional layers and 3D maximum pooling layers, the spatial dimensions of the image gradually decrease, while the number of feature channels (i.e., the number of filters) gradually increases, forming a series of higher-level feature representations that capture the complex structures and patterns in the image. The global average pooling layer compresses the spatial dimensions of each feature map into a single value, that is, calculates the average value of each feature map. This step generates a fixed-length feature vector that contains the global information of the entire 3D image. The final fully connected layer further processes the feature vector output by the global average pooling layer, learns the complex relationship between features, and generates the final 128-dimensional feature vector, i.e., image features.

[0155] 3D CNN needs to be pre-trained on a large number of labeled 3D medical image datasets to learn the key features in medical images. This process can be achieved through an autoencoder structure or training for specific tasks (such as image segmentation or classification). The advantage of the pre-trained model is that it can capture complex patterns and relationships in 3D space and learn highly abstract features, which are essential for representing various medical-related features.

[0156] In S62, principal component analysis (PCA) is used to compress the extracted text features, numerical features and image features respectively. This process specifically uses the PCA principle to determine the main change direction of the data, namely the principal components. PCA converts the original possibly correlated variables into a set of linearly uncorrelated variables through orthogonal transformation. These variables are called principal components.

[0157] When processing several original feature vectors such as text features, numerical features, and image features, principal component analysis (PCA) is used to compress them so that the compressed features (referred to as compressed features) can be transmitted to the operation center. Specifically, the PCA parameter `n_components` is set to 0.95, which means that enough principal components are retained to explain 95% of the variance of the original data. PCA will automatically calculate and retain the necessary number of principal components to ensure that the compressed features can meet this variance explanation rate. The compression ratio is calculated based on the ratio of the number of principal components retained to the original feature vector. For example, compressing a 128-dimensional original feature vector to 64 dimensions has a compression ratio of 0.5. The degree of compression of PCA will be automatically adjusted according to the actual variance distribution of the original feature vector. Therefore, even if the same variance retention rate is set for different types of original feature vectors, the final compressed feature dimensions may be different. After these compressed features are processed locally, they are transmitted to the computing center to prepare for subsequent data analysis. This can significantly reduce data dimensions and storage and transmission costs, making it particularly suitable for use in situations where bandwidth is limited. It can also remove noise and redundant information, retain the most important features, and improve the efficiency and accuracy of subsequent health tests. This local computing method can be considered edge computing and plays a certain role in privacy protection. In order to further protect patient privacy, additional encryption measures are taken during the transmission process to ensure data security. This method can reduce the amount of data transmission while retaining key information, providing support for efficient data analysis.

[0158] When implementing this solution, several key factors need to be considered to ensure the efficiency and effectiveness of the entire process. First, since the feature extraction process of 3D CNN can be very computationally intensive, GPU can be used to accelerate this step to improve computational efficiency. Second, considering that the computational complexity of PCA is proportional to the number of features and the amount of data, for large-scale data, incremental PCA is used to process it. Before applying PCA, the features extracted by 3D CNN are standardized to ensure the consistency of the proportions of each feature, which is crucial to the effect of PCA.

[0159] In order to adapt the model to potential changes in data distribution, the 3DCNN is regularly retrained with the latest medical data and the PCA is updated to keep it up to date and improve its adaptability to new data. To improve the interpretability of the model, interpretable AI techniques such as layer-wise relevance propagation are used to better understand the features extracted by CNN and their contribution to the final decision.

[0160] In S63, the computing center can be in the cloud, and can use more powerful computing resources to perform complex anomaly detection algorithms. The computing center decompresses the received compressed features to obtain a number of text features, numerical features, and image features. Figure 6 The flowchart of the three-dimensional image construction method shown in the figure shows that after all image features are obtained, a three-dimensional image is constructed based on these image features, which specifically includes the following steps:

[0161] S631: Determine whether the number of image features is equal to 1; if so, execute S632; if not, execute S634;

[0162] S632: construct a comprehensive three-dimensional image based on the text features, numerical features and image features, and execute S633;

[0163] S633: Perform health checks on patients based on comprehensive three-dimensional images;

[0164] S634: assigning a fusion level to each image feature, where the fusion level is 0;

[0165] S635: Determine whether the number of image features is an odd number; if so, execute S636; if not (i.e., an even number), execute S637;

[0166] S636: adding 1 to the fusion level of any image feature, and executing S637;

[0167] S637: Fuse image features of the same fusion level in pairs to obtain several new image features, and return to S631.

[0168] Note that when S634 is executed again, a fusion level of 0 will be reassigned to each image feature.

[0169] Through the above method, when there are multiple decompressed image features, they are fused and a comprehensive image feature can be finally obtained. For example, the image features extracted from the CT scan image and the MRI image are fused, so that the text features and numerical features can be used to assist the comprehensive image features in the subsequent construction of a three-dimensional image, and a comprehensive three-dimensional image integrating multiple data can be obtained. When there is only one decompressed image feature, a comprehensive three-dimensional image is constructed directly based on the text features, numerical features and image features.

[0170] Among them, the pairwise fusion of image features is illustrated by an example: there are 4 medical images, and correspondingly, there are also 4 image features, namely A, B, C, and D. A and B are fused to obtain a new image feature E, C and D are fused to obtain a new image feature F, and E and F are fused to obtain the final image feature G.

[0171] In the technical solution of fusing two image features, a distance measurement method (Euclidean distance or cosine similarity) is first used to calculate the distance between each feature element of the two image features to be fused. The distance between the feature elements of the two image features constitutes a cost matrix, which reflects the similarity between each pixel or region in the medical image represented by the two image features. The Hungarian algorithm (also known as the Kuhn-Munkres algorithm) is used to solve the discrete process matching problem. The algorithm can find the minimum cost match in the cost matrix to minimize information loss. According to the result of the minimum cost match, the two image features are fused by a weighted average method. This method takes into account the importance of each feature to achieve more effective information merging. In addition, fusion techniques such as wavelet transform or principal component analysis (PCA) can be used to further improve the quality of fusion and ensure that key image information can be retained at different scales. Through this series of steps, the features of multimodal medical images can be effectively combined to provide more comprehensive and accurate data support for subsequent diagnosis.

[0172] In S632 and S633, data fusion technology is used to fuse text features, numerical features and image features, and a comprehensive three-dimensional image is constructed based on the fused features. The generated comprehensive three-dimensional image is post-processed, such as smoothing, denoising or enhancing, to improve the image quality. Medical staff can use three-dimensional visualization tools to view and analyze the comprehensive three-dimensional image for further diagnosis or analysis of the patient. The comprehensive three-dimensional image is analyzed based on a preset detection model to achieve health detection of the patient. Among them, the detection model can use a deep learning model, such as a U-Net or ResNet variant, which can effectively capture the contextual information in the image and accurately locate the target area through its unique U-shaped structure. A large number of labeled three-dimensional medical image data sets are used to pre-train the selected detection model. These data can be images of different modalities, such as CT, MRI or PET scans, or multi-modal fusion images, combined with corresponding annotation information, to guide the detection model to learn to distinguish different tissues and structures. On the comprehensive three-dimensional image, the pre-trained detection model is used to extract features. These features can represent the key information in the image, such as shape, texture, and position. For segmentation tasks, the detection model will learn to distinguish different tissues and structures, such as healthy tissue and diseased areas. For classification tasks, the detection model will classify the entire image or specific areas in the image to determine whether there are abnormalities. The detection model can also use multimodal deep learning models, such as multimodal transformers, which can comprehensively analyze all available information to achieve deep health detection of patients. Regularly retrain and update the detection model with new medical data to adapt to changes in data distribution and ensure the accuracy and robustness of the detection model.

[0173] Set the anomaly score threshold. The detection model scores the patient according to the detected abnormalities and outputs the anomaly score. If the anomaly score is greater than the anomaly score threshold, it indicates that an abnormality has occurred. If the anomaly score is not greater than the anomaly score threshold, it indicates that no abnormality has occurred.

[0174] In S7, when it is determined that an abnormality occurs, an early warning message is generated for medical staff to refer to and provide treatment guidance to the patient, and the follow-up task is completed. The early warning message includes at least one of the patient's basic information, abnormality type, abnormality location, abnormality score, abnormality degree (determined according to the abnormality score), and recommended measures. According to the degree of abnormality, different notification strategies are used to send early warning information to the terminal device of the medical staff, such as text messages, emails, system internal messages, etc.

[0175] In one possible implementation, when it is determined that an abnormality occurs, the follow-up task is re-pushed in accordance with the first possible implementation method of S2 described above.

[0176] In another possible implementation, when an abnormality is determined to have occurred, a new severity level is generated based on the abnormality score, the patient list is updated based on the new severity level, and then the follow-up task is re-pushed in the manner of the second possible implementation of S2 described above.

[0177] In S8, when no abnormality occurs, it means that the follow-up task is completed. The medical staff determines whether the patient has recovered based on the follow-up situation. If not, the follow-up task is pushed again in accordance with the first feasible solution of S2 above. Alternatively, the medical staff determines whether the patient's severity level remains unchanged or decreases by 1 based on the follow-up situation. If it decreases by 1, the patient list is updated. If it remains unchanged, there is no need to update the patient list, and then the follow-up task is pushed again in accordance with the second feasible solution of S2 above. If the patient recovers, the medical staff clicks the "Terminate Follow-up" button on the follow-up system interface through a mouse or other device to trigger a follow-up end request for the patient. In response to the follow-up end request, the basic information of the recovered patient is deleted from the patient list, and the follow-up task for the patient is no longer pushed.

[0178] In S7 and S8, when each follow-up task is completed, the follow-up task is automatically saved in the historical follow-up list. The follow-up system supports viewing and exporting follow-up history records.

[0179] Optionally, provide a patient mobile application to support the upload of follow-up records, diagnostic data, electronic medical records, etc. Patients can view follow-up plans, receive health advice, etc. Incentive mechanisms can also be issued on the patient mobile terminal to promote patients' active participation in medical management. Regularly maintain the follow-up system and the patient mobile terminal system to ensure stable operation and continuously optimize system functions based on user feedback.

[0180] In summary, the intelligent follow-up system disclosed in the present invention can improve the quality and efficiency of medical services. The system reduces the workload of medical staff through automation and intelligent means, achieves a significant improvement in follow-up efficiency, and can formulate personalized follow-up plans based on the patient's specific condition and treatment history to achieve precise management. At the same time, the system will integrate the patient's medical information, including medical records, examination results and treatment records, build a complete medical file, and promote data sharing between different medical institutions. In addition, the system will optimize the patient experience and provide convenient and humane services. The system also assumes the role of preventive health care and health education to help patients maintain their own health. The system can serve a variety of medical institutions such as hospitals, clinics and community medical centers, especially in chronic disease management, rehabilitation treatment and postoperative follow-up scenarios, to provide patients with comprehensive and effective medical services.

[0181] Reference Figure 7 The present disclosure provides a patient post-diagnosis intelligent follow-up system, comprising:

[0182] A creation module 101 is used to create a staff list and a patient list;

[0183] A determination module 102, for determining a push time based on a patient list;

[0184] A selection module 103, for selecting the best medical staff based on the staff list and the patient list when the push time arrives;

[0185] A push module 104 is used to push the follow-up task to the terminal device of the best medical staff;

[0186] The acquisition module 105 is used to obtain the medical data of the patient in response to the medical staff's follow-up request for the patient;

[0187] A detection module 106, for performing health detection on a patient based on medical data;

[0188] The generating module 107 is used to generate warning information if an abnormality occurs, provide treatment guidance to the patient based on the warning information, and re-send the follow-up task.

[0189] The various variations and specific examples of the above-mentioned intelligent follow-up method for patients after diagnosis are also applicable to the intelligent follow-up system for patients after diagnosis provided in the present disclosure. Through the above-mentioned detailed description of the intelligent follow-up method for patients after diagnosis, those skilled in the art can clearly know the implementation method of the intelligent follow-up system for patients after diagnosis. For the sake of brevity of the specification, it will not be described in detail here.

[0190] The computer device according to the embodiment of the present disclosure includes a memory and a processor. The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0192] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory, so that the computer device performs all or part of the steps of the patient post-diagnosis intelligent follow-up method of each embodiment of the present disclosure.

[0193] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present disclosure.

[0194] like Figure 8 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure is shown, which is a schematic diagram of the structure of a computer device suitable for implementing the embodiment of the present disclosure. Figure 8 The computer device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0195] like Figure 8 As shown, the computer device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage device into a random access memory (RAM). In the RAM, various programs and data required for the operation of the computer device are also stored. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0196] Typically, the following devices can be connected to the I / O interface: input devices such as sensors or visual information acquisition devices; output devices such as display screens; storage devices such as tapes, hard disks, etc.; and communication devices. The communication device can allow the computer device to communicate with other devices (such as edge computing devices) wirelessly or by wire to exchange data. Figure 8A computer device having various devices is shown, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0197] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the patient post-diagnosis intelligent follow-up method of the embodiment of the present disclosure are executed.

[0198] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0200] According to the computer-readable storage medium of the embodiment of the present disclosure, non-transitory computer-readable instructions are stored thereon. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the patient post-diagnosis intelligent follow-up method of each embodiment of the present disclosure are executed.

[0201] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0202] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0203] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0204] In the present disclosure, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The block diagrams of the devices, devices, equipment, and systems involved in the present disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0205] Additionally, as used herein, "or" used in a list of items beginning with "at least one" indicates a separate list, so that, for example, a list of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not mean that the example described is preferred or better than other examples.

[0206] It should also be noted that in the system and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0207] YY+242248P

[0208] Various changes, substitutions, and modifications of the techniques described herein may be made without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of the present disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of events, means, methods, and actions described above. Currently existing or later to be developed processes, machines, manufactures, compositions of events, means, methods, or actions that perform substantially the same functions or achieve substantially the same results as the corresponding aspects described herein may be utilized. Thus, the appended claims include such processes, machines, manufactures, compositions of events, means, methods, or actions within their scope.

[0209] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0210] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for intelligent follow-up of patients after diagnosis, characterized in that: include: Create staff and patient lists; Determining a push time based on the patient list; When the push time arrives, the best medical staff is selected based on the employee list and the patient list; Pushing the follow-up task to the terminal device of the optimal medical staff; Responding to a medical staff follow-up request for a patient, obtaining the patient's medical data; Conduct health checks on patients based on medical data; If an abnormality occurs, an early warning message is generated, treatment guidance is given to the patient based on the early warning message, and a follow-up task is re-sent.

2. The intelligent follow-up method for patients after diagnosis according to claim 1, characterized in that: The calculation formula of the push time is: T 推送 =T i就诊 +X*T 基础 -T 浮动 ; Where, T 推送 is the push time; i is the patient's visit number; T i就诊 is the patient's consultation time; the patient list includes the patient's consultation number and consultation time; X is the current follow-up number; T 基础 is the preset base date; T 浮动 The default floating date.

3. The intelligent follow-up method for patients after diagnosis according to claim 1, characterized in that: The calculation formula of the push time is: T 推送 =T i就诊 +X*(T 基础 -T j )-T 浮动 ; Where, T 推送 is the push time; i is the patient's visit number; T i就诊 is the patient's visit time; X is the current follow-up number; T 基础 is the preset base date; j is the severity level; T j T is the advancement time of the severity level; 浮动 It is a preset floating date; the patient list includes the patient's visit number, visit time and severity level.

4. The intelligent follow-up method for patients after diagnosis according to claim 1, characterized in that: Based on the employee list and the patient list, obtaining a comprehensive matching score between each medical staff member and the patient, and selecting the medical staff member with the highest comprehensive matching score as the best medical staff member; The expression of the comprehensive matching score is as follows: FH=w S ·S+w W ·W+w H ·H+w P ·P+w F ·F; Wherein, FH is the comprehensive matching score; w S 、w W 、w H 、w P 、w F is the preset weight; S is the expertise matching score; W is the workload score; H is the historical interaction record score; P is the patient demand score; F is the familiarity score; Where n is the total number of disease types, 1 <k≤n;p k is the kth element in the medical staff expertise vector, indicating whether the medical staff has expertise in treating the kth disease. If so, p k =1, if not, then p k =0;d k is the kth element in the patient's disease type vector, indicating whether the patient suffers from the kth disease. If so, d k =1, if not suffering from, then d k =0; Among them, Z is the current follow-up workload; Zm is the historical maximum follow-up workload; a, b, and c are all preset adjustment parameters; Among them, λ is the initial interaction score; e is a natural constant; t is the number of days experienced by the medical staff and the patient from the last actual follow-up time to the current time; μ is the decay parameter; Among them, β0 is the basic demand score; β1 is the degree of influence of the patient's age on the demand; L is the patient's age; β2 is the degree of influence of the patient's disease type on the demand; Where V is the current total number of times in the statistical period; x v is the frequency of follow-up visits of medical staff to patients during the vth statistical period; is the average follow-up frequency in all statistical periods; v is the patient satisfaction in the vth statistical period; It is the average of the patients' satisfaction during all statistical periods.

5. The intelligent follow-up method for patients after diagnosis according to claim 1, characterized in that: The step of obtaining the patient's medical data in response to the medical staff's request for a follow-up visit to the patient includes: In response to a follow-up request from a medical staff to a patient, providing the medical staff with a follow-up questionnaire template; Responding to a submission request from a medical staff member, obtaining a completed follow-up questionnaire template; Extract information from the completed follow-up questionnaire template to obtain the patient's current condition data; YY+242248P Access to clinical data of patients across several healthcare systems; Dividing the current condition data and the clinical data according to data categories to obtain a number of text data, numerical data and medical images; The plurality of text data, numerical data and medical images constitute the medical data.

6. The intelligent follow-up method for patients after diagnosis according to claim 5, characterized in that: The patient's health check based on medical data includes: Performing local feature extraction on the plurality of text data, numerical data and medical images respectively to obtain a plurality of text features, numerical features and image features; Compressing the plurality of text features, numerical features and image features respectively, and transmitting the compressed features to a computing center; The compressed features are decompressed by the computing center, and health checks are performed on the patient based on the decompressed text features, numerical features, and image features.

7. The intelligent follow-up method for patients after diagnosis according to claim 1, characterized in that: The health check of the patient based on the decompressed text features, numerical features and image features includes: If the number of the image features is equal to 1, constructing a comprehensive three-dimensional image based on the text features, the numerical features and the image features; Based on the comprehensive three-dimensional image, performing health examination on the patient; If the number of the image features is not equal to 1, a fusion level is assigned to each image feature, and the fusion level is 0; Determining whether the number of the image features is odd; If it is an odd number, the fusion level of any image feature is increased by 1, and image features with the same fusion level are fused in pairs to obtain a number of new image features, and it is re-determined whether the number of the image features is equal to 1; If it is not an odd number, image features of the same fusion level are fused in pairs to obtain a number of new image features, and it is re-determined whether the number of the image features is equal to 1.

8. A computer device, characterized in that: The computer device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method of claim 1. Intelligent follow-up method after diagnosis.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which are used to enable a computer to execute the patient post-diagnosis intelligent follow-up method described in any one of claims 1-7.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images