Hierarchical chronic disease management and referral system based on artificial intelligence

Abstract

The invention discloses a hierarchical chronic disease management and referral system based on artificial intelligence, and belongs to the technical field of intelligent medical treatment, and the system comprises the steps: setting upstream and downstream referral paths composed of diagnosis and treatment institutions, binding responsible doctors corresponding to chronic diseases for the diagnosis and treatment institutions of each level, and forming a referral relation mapping table; predicting a chronic disease dynamic risk based on the multi-source and multi-modal data of the target crowd, classifying the target crowd, generating a chronic disease management suggestion for each type of crowd, and pushing the chronic disease management suggestion to a chronic disease doctor for contract signing management; generating referral suggestions based on the chronic disease dynamic risk changes of each type of people and pushing the referral suggestions to chronic disease doctors, and the chronic disease doctors determining the referral suggestions, setting referral information based on upstream and downstream referral paths and a referral relation mapping table and synchronously notifying target diagnosis and treatment institutions and patients; performance excitation is performed on whether a chronic disease doctor completes contract signing of a chronic disease patient, implementation of graded referral treatment and medical cost saving, so that the medical service precision can be improved, the diagnosis and treatment efficiency is improved, and medical cost throttling is promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent healthcare, and particularly relates to a hierarchical chronic disease management and referral system based on artificial intelligence. Background Art

[0002] Hierarchical diagnosis and treatment and referral play an important role in the healthcare model centered on health. At present, in the process of referral and hierarchical diagnosis and treatment, patient referral and stratification mainly rely on the manual judgment of medical staff based on guidelines and experience and patient-initiated referral. Due to the demand for high-quality medical services, patients also tend to seek medical treatment in higher-level hospitals during the stable period of chronic diseases, resulting in a heavy medical burden on higher-level hospitals and insufficient utilization of primary medical services.

[0003] In recent years, artificial intelligence technology and big data have developed rapidly, and the artificial intelligence (AI) industry has continuously given birth to new business forms through deep integration in various fields. With its superior performance such as big data-assisted diagnosis, intelligent decision-making support, real-time information sharing, resource allocation prediction, and automated processing, AI helps to promote the reform and optimization of medical hierarchical diagnosis and treatment and referral. Therefore, there is an urgent need for a hierarchical management and referral system based on artificial intelligence. Summary of the Invention

[0004] In view of the above, the purpose of the present invention is to provide a hierarchical chronic disease management and referral system based on artificial intelligence, which establishes a hierarchical diagnosis and treatment system through artificial intelligence technology, classifies and stratifies the target management population, constructs an upstream and downstream referral chain for medical institutions and medical teams, conducts proactive management and proactive referral based on artificial intelligence risk prediction, constructs a reverse incentive mechanism, and optimizes the allocation of medical resources and medical insurance management.

[0005] To achieve the above object of the invention, an embodiment provides a hierarchical chronic disease management and referral system based on artificial intelligence, including: A target population classification and signing management module, which is used to predict the dynamic risk of chronic diseases based on multi-source and multi-modal data of the target population, classify the target population, generate chronic disease management suggestions for each category of population, and push them to chronic disease doctors for signing management; A hierarchical referral path setting module, which is used to provide a hierarchical referral path setting function, set the upstream and downstream referral paths composed of medical institutions based on the referral path setting function, bind the responsible doctors corresponding to chronic diseases to each level of medical institutions, and form a referral relationship mapping table; A proactive referral module, which is used to generate referral suggestions based on the dynamic risk changes of chronic diseases of each category of population and push them to chronic disease doctors. The chronic disease doctors determine the referral suggestions, set referral information based on the upstream and downstream referral paths and the referral relationship mapping table, and synchronously notify the target medical institutions and patients of the referral; A reverse incentive module, which is used to perform performance incentives for chronic disease doctors on whether they complete the signing of chronic disease patients, implement hierarchical referral, and save medical expenses.

[0006] Preferably, in the target population classification and signing management module, the dynamic risk of chronic diseases is predicted based on multi-source and multi-modal data of the target population, and the target population is classified. Chronic disease management suggestions are generated for each type of population and pushed to chronic disease doctors for signing management, including: Using a chronic disease prediction model to predict the dynamic risk of chronic diseases based on multi-source and multi-modal data of the target population; Target populations with chronic disease dynamic risk prediction results lower than the first threshold are classified as healthy populations. For healthy populations, follow-up reminders and follow-up items are regularly pushed to chronic disease doctors and patients, the follow-up item data of healthy populations is actively tracked, and the prediction of the progression to high-risk populations is observed; Target populations with chronic disease dynamic risk prediction results exceeding the second threshold are classified as high-risk populations, and chronic disease doctors are signed for high-risk populations on the principle of signing as many as willing. The follow-up frequency is increased according to the personalized disease risks of high-risk individuals, the follow-up frequency data of high-risk populations is actively tracked, and after the expected follow-up interval exceeding a certain number of times is not completed, a warning value will be triggered and the chronic disease doctor will be actively reminded to pay attention to high-risk individuals; Populations with clinically diagnosed diseases are classified as diagnosed populations, and populations with actual diseases that have not been diagnosed by doctors are classified as populations to be managed. For diagnosed populations and populations to be managed, all are required to sign chronic disease doctors, and signing and long-term management are achieved through methods such as telephone mobilization, APP reminders, tracking medical records, and health education during high-risk periods.

[0007] Preferably, using a chronic disease prediction model to predict the dynamic risk of chronic diseases based on multi-source and multi-modal data of the target population, including: First, based on three types of structured multi-modal data of the target population's historical diagnosis data, biochemical indicators, and follow-up information, a multi-modal health trajectory in the form of a time series is constructed; Secondly, the chronic disease prediction model adopts a fusion method of a dynamic risk coding mechanism and an attention mechanism. The multi-modal health trajectory data is input into the chronic disease prediction model according to a certain time window. First, it is encoded, and the attention mechanism is used for the encoded features to capture the non-linear change characteristics of the key variables concerned in the data under different time windows. The convolutional neural network and long short-term memory network adopted by the multi-task learning head are used to fuse the non-linear change characteristics of each key variable, and the dynamic risk of chronic diseases and the trend of health status in the future time period are predicted based on the fusion result. Among them, the dynamic risk of chronic diseases is represented in the form of probability, and the larger the value, the higher the risk; An adaptive time decay function is also embedded in the chronic disease prediction model, introducing a mechanism for decreasing the weight of historical data in the long term. Different decay weights are set for the input data within the time window according to the distance from the prediction time to adjust the influence degree on the prediction result. Specifically, the input data farther away from the prediction time is set with a larger decay weight.

[0008] Preferably, for the long-term management of the diagnosed population and the population to be managed, it includes: Implement hierarchical management according to the progress of chronic diseases. Specifically, based on the diagnosis and treatment consensus and guidelines of the clinical multidisciplinary doctor team as prior knowledge, for patients with different chronic disease processes, judge the cardiovascular disease risk stratification according to the cardiovascular disease risk stratification model, where the stratification includes low risk, medium risk, high risk, and very high risk, and carry out follow-up frequencies and follow-up items with different requirements according to the stratification results; Among them, judging the disease progression risk level according to the risk stratification model includes: (1) Construct a knowledge graph based on the consensus of clinicians to clarify the risk factors of chronic diseases and their importance; (2) Using the historical patient database, with the physiological indicators of the patients as independent variable data and each type of chronic disease risk factor judged by complications as the dependent variable, construct a logistic regression model to obtain the regression coefficients of each indicator for each type of chronic disease risk factor, and then normalize the absolute values of the regression coefficients to obtain the contribution ratio of each indicator to each type of chronic disease risk factor; (3) Calculate the real-time risk score of the patient using the weighted summation method based on the importance of each type of chronic disease risk factor and the contribution ratio of each indicator to each type of chronic disease risk factor; (4) Apply the entropy weight method to dynamically analyze the risk score distribution of all patients, and set the dynamic threshold of risk stratification according to the risk score distribution; (5) Real-time collect the physiological indicator data of the patient individual and calculate the real-time risk score, and accurately stratify the patient to the corresponding risk level according to the dynamic threshold of risk stratification.

[0009] Preferably, when the hierarchical referral path setting module is logged in by a super administrator with high permissions, it is used to set the upstream and downstream referral paths, and the set upstream and downstream referral paths include: community health service station at level 1 → community health service center at level 2 → district-owned hospital at level 3 → tertiary hospital at level 4; and separately implement referral hospitals and target responsible doctor teams for high-risk chronic disease populations; When the hierarchical referral path setting module is logged in by a chronic disease doctor, fill in the list of superior referral doctors, and the nurse fills in the assistance management path.

[0010] Preferably, in the active referral module, referral suggestions are generated based on the dynamic risk changes of chronic diseases for each type of population and pushed to chronic disease doctors, including: When the diagnosed population or the population to be managed is determined to be at continuous high risk based on the dynamic risk changes of chronic diseases, or the follow-up is not completed, it is determined to be a high-risk situation. When the high-risk population is continuously at high risk, it will also be judged as a high-risk situation; Use the referral recommendation generation model to generate referral recommendations for the population in high-risk situations and push them to chronic disease doctors. The referral recommendations include a patient risk analysis report containing an index trend chart and risk assessment, recommended referral hospitals and doctors matched according to the pre-set hierarchical referral path, and a recommended referral time dynamically recommended based on the source data of the superior hospital.

[0011] Preferably, the process of the referral recommendation generation model generating referral recommendations for the population in high-risk situations is as follows: (1) Calculate the medical gap index according to the level difference and disease treatment capabilities between the patient's current medical institution and the target medical institution, where the medical gap index is the product of the level gap between the two medical institutions and the disease treatment capabilities; (2) Real-time monitor the patient's key physiological indicators and complication risk scores, and input the monitoring data combined with the medical gap index into the gated recurrent neural network to output the emergency score in real time, and determine high-risk patients based on the emergency score; (3) Input the emergency score into the asymmetric priority algorithm. For the emergency score of high-risk patients, for every given increase in the score, the queuing priority weight will correspondingly increase by a given multiple, thereby updating the priority order of the patient queue; (4) Collect the medical resource status of the target hospital to generate a resource heat map. Based on the updated patient sequence with priorities, input the patient's priority and the hospital resource heat map into a multi-objective integer programming model. The multi-objective integer programming model takes the patient's priority as the input variable, and at the same time takes the medical resource status of the target hospital as the constraint condition to construct the objective function. The objective function aims to maximize the sum of the products of the patient's priority weight and the medical resource satisfaction, and then uses integer programming to solve the model to obtain the optimal referral path, realize the best match between patient needs and medical resources, and determine the optimal referral path between patient needs and medical resources; (5) Automatically generate a referral report, including the optimal referral path, the patient's real-time emergency score, an overview of the medical resource status of the referral hospital, the expected reduction in risk after referral, and the reasons for the referral recommendation.

[0012] Preferably, in the active referral module, the chronic disease doctor determines the referral recommendation and sets the referral information based on the upstream and downstream referral paths and the referral relationship mapping table and synchronously notifies the target diagnosis and treatment institution and the patient of the referral, including: After receiving the referral recommendation and reviewing the prediction basis, the chronic disease doctor selects to accept the referral recommendation or make manual modifications to confirm and generate the final referral information, that is, the final referral information is generated by setting information according to the upstream and downstream referral paths and the referral relationship mapping table, where manual modification takes precedence over the generated referral recommendation; When the chronic disease doctor rejects the referral recommendation, the reason must also be filled in manually and an electronic signature must be made.

[0013] Preferably, the active referral module also records in real time when the target diagnosis and treatment institution accepts the referral and generates a treatment plan, and synchronizes the treatment plan to the chronic disease doctor.

[0014] Preferably, in the reverse incentive module, performance incentives are given to chronic disease doctors for whether they complete the signing of chronic disease patients, implement hierarchical referrals, and save medical expenses, including: According to the specific referral situation recorded in the system, the model prediction results, whether there is a referral and modification record, the follow-up completion situation, and the referral effectiveness, the referral acceptance rate and the medical expense savings amount are statistically calculated, and the performance incentive amount is calculated in combination with the basic salary, the signing rate, the referral acceptance rate, and the medical expense savings amount.

[0015] Compared with the prior art, the beneficial effects of the present invention at least include: The hierarchical and classified chronic disease management and referral system based on artificial intelligence provided by the present invention improves the efficiency of chronic disease management, optimizes the allocation of medical resources, and saves medical expense expenditures through various models constructed by artificial intelligence. Specifically, the target population is classified through a hierarchical diagnosis and treatment system, and upstream and downstream referrals are carried out according to risk prediction and the set hierarchical referral path. The manually driven active screening and active referral mechanism can efficiently identify high-risk patients and automatically generate management and referral recommendations to ensure timely intervention. Through the reverse incentive mechanism, the enthusiasm of medical staff is stimulated. This system can improve the accuracy of medical services, improve the diagnosis and treatment efficiency, and promote the reduction of medical expenses. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a schematic structural diagram of a hierarchical and classified chronic disease management and referral system based on artificial intelligence provided by the embodiment; Figure 2 is a flow chart of hierarchical and classified chronic disease management and referral provided by the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0019] The inventive concept of the present invention is as follows: An artificial intelligence-based hierarchical and classified chronic disease management and referral system is provided in an embodiment of the present invention. Guided by artificial intelligence, a chronic disease prediction model is used to predict the risks of chronic disease onset and disease progression of an individual, classify the individual into high-risk and low-risk groups, and with two-way and dual-pathways as the core and manual as the final review, active recommendations are given based on the risk prediction results, and the review by chronic disease doctors is carried out to make a final decision, and corresponding active management and active referral are carried out. Among them, two-way refers to upward referral and downward referral to form a complete upstream and downstream chain of diagnosis and treatment, and promote the reasonable flow of medical resources and social medical insurance. Dual-pathways refer to the pathway of the chronic disease population (confirmed population + population to be managed) and the pathway of the high-risk population.

[0020] The artificial intelligence-based hierarchical and classified chronic disease management and referral system provided by the embodiment, to achieve precise health and chronic disease management, establishes a hierarchical and classified diagnosis and treatment system based on risk prediction, which is specifically reflected in the hierarchical referral pathway and the classified management of the target population, ensuring that each type of population receives appropriate medical resource allocation; it also has an artificial intelligence-driven active management and active referral mechanism, and artificial intelligence will play a core role in the hierarchical and classified diagnosis and treatment management of chronic diseases, actively screening different classified populations and including them in management, and actively making referral recommendations, using artificial intelligence to assist in improving the screening rate, diagnosis rate and treatment rate of chronic diseases, assisting in the rational allocation of medical resources and assisting in the diversion of medical insurance; it also has a reverse incentive mechanism to stimulate the initiative and enthusiasm of the medical staff team and continuously optimize the referral pathway and management.

[0021] As Figure 1 shown, the artificial intelligence-based hierarchical and classified chronic disease management and referral system 10 provided by the embodiment includes a target population classification and contract management module 11, a hierarchical referral pathway setting module 12, an active referral module 13, and a reverse incentive module 14.

[0022] In the embodiment, the target population classification and contract management module 11 is used to divide the target management population included in the database into a confirmed population, a population to be managed, a high-risk population and a healthy population, and perform hierarchical management. Among them, the confirmed population refers to the population that has been diagnosed with chronic diseases by a doctor, the population to be managed refers to the population that actually meets the chronic disease diagnosis criteria but has not been diagnosed by a doctor, the high-risk population refers to the population that does not currently meet the chronic disease diagnosis criteria but is identified as high-risk of onset by the chronic disease prediction model, and the healthy population refers to the population that is identified as low-risk of onset by the chronic disease prediction model.

[0023] In patient information entry, the chronic disease doctor assists the patient in logging in for the first time, filling in personal information, and selecting whether to sign and bind. For the diagnosed population and the population to be managed, ensure that they sign and bind with the chronic disease doctor; mobilize the high-risk population to sign and bind as much as possible. Predict the dynamic risk of chronic diseases through an artificial intelligence-based chronic disease prediction model, and further refine the disease progression risk levels of the diagnosed population and the population to be managed to facilitate further management.

[0024] Specifically, predict the dynamic risk of chronic diseases based on the multi-source and multi-modal data of the target population and classify the target population, generate chronic disease management suggestions for each type of population and push them to the chronic disease doctor for signing management. Different types of populations adopt different signing management models and chronic disease management approaches, including: Use the chronic disease prediction model to predict the dynamic risk of chronic diseases based on the multi-source and multi-modal data of the target population. Specifically, it includes: First, based on the three types of structured multi-modal data of the historical diagnosis data, biochemical indicators, and follow-up information of the target population, construct a multi-modal health trajectory in the form of a time series. This multi-modal health trajectory reflects the dynamic change trend of the patient's health status through the time series; among them, the historical diagnosis data includes the diagnosis time and degree of hypertension, diabetes, etc., and the biochemical indicator data includes the regular test results of fasting blood glucose, glycated hemoglobin, total cholesterol, etc. Secondly, the chronic disease prediction model adopts a fusion method of a dynamic risk coding mechanism and an attention mechanism, that is, the multi-modal health trajectory data is input into the chronic disease prediction model according to a certain time window. First, it is encoded and the attention mechanism is used for the encoded features to capture the non-linear change features of the key variables concerned in the data under different time windows, more effectively identifying the key nodes of risk changes (for example, the fasting blood glucose suddenly rises from 5.5 mmol / L to 7.8 mmol / L). The non-linear change features of each key variable are fused through the convolutional neural network (CNN) and long short-term memory network (LSTM) adopted by the multi-task learning head, and the dynamic risk of chronic diseases and the health status trend in the future time period are predicted based on the fusion result. Among them, the dynamic risk of chronic diseases is represented in the form of probability, and the larger the value, the higher the risk. The health status trend includes, for example, that the glycated hemoglobin may reach 8.5% in the next few months. It should also be noted that an adaptive time decay function is embedded in the chronic disease prediction model, introducing a weight decay mechanism for the long-term historical data. Specifically, different decay weights are set for the input data within the time window according to the distance from the prediction time to adjust the influence degree on the prediction result. Specifically, the input data farther away from the prediction time is set with a larger decay weight. For example, the weight of the input data in the recent 6 months is 1, the weight of the input data from 6 to 12 months is 0.6, and the weight of the input data more than 12 months is reduced to 0.3.

[0025] For healthy populations: Target populations with chronic disease dynamic risk prediction results lower than the first threshold are classified as healthy populations. Follow-up reminders and follow-up items are regularly pushed to chronic disease doctors and patients via PC and computer terminals. For example, for hypertension, the follow-up frequency is set at once a year. At the same time, actively track the follow-up item data of healthy populations and observe the prediction of those who progress to high-risk populations.

[0026] For high-risk populations: Target populations with chronic disease dynamic risk prediction results exceeding the second threshold are classified as high-risk populations, where the second threshold is greater than the first threshold. For high-risk populations, the requirement for signing contracts with chronic disease doctors is to sign as many as willing, and actively mobilize, that is, sign chronic disease doctors for high-risk populations on the principle of signing as many as willing. The classification of high-risk populations is no longer based on traditional guidelines but on the personalized individual risks of patients. Increase the follow-up frequency based on the personalized disease risks of high-risk individuals. For example, for patients with both diabetes and chronic kidney disease, the actual AI criteria for including them in the high-risk population of hypertension will be lower than the 120 - 130 mmHg systolic blood pressure or 80 - 89 mmHg diastolic blood pressure in the guidelines. The expected follow-up frequency for this patient will be increased to 2 - 3 times a year. At the same time, actively track the follow-up frequency data of high-risk populations. After a certain number of expected follow-up intervals are not completed, a warning value will be triggered and the chronic disease doctor will be actively reminded to pay attention to high-risk individuals.

[0027] For confirmed populations and populations to be managed: Clinically diagnosed diseased populations are classified as confirmed populations, and populations with actual diseases not yet diagnosed by doctors are classified as populations to be managed. For confirmed populations and populations to be managed, it is required that all sign contracts with chronic disease doctors, and achieve signing and long-term management through methods such as telephone mobilization, APP reminders, tracking medical records, and health education during high-risk periods.

[0028] Among them, long-term management includes: hierarchical management according to the progress of chronic diseases. Specifically, based on the clinical multidisciplinary doctor team diagnosis and treatment consensus and guidelines as prior knowledge, for patients with different chronic disease progressions (for example, patients with grade 1 / 2 / 3 hypertension), judge the disease progression risk level according to the risk stratification model. The stratified levels include low risk (low risk), medium risk (medium risk), high risk (high risk), and very high risk (very high risk), and carry out follow-up frequencies and follow-up items with different requirements according to the stratification results.

[0029] Among them, judging the disease progression risk level according to the risk stratification model includes: (1) Construct a knowledge graph based on clinical doctor consensus to clarify the risk factors of chronic diseases and their importance levels (for example, the importance score of grade 1 hypertension is 0.3, grade 2 is 0.6, grade 3 is 0.9, and diabetes is 0.7); (2)Using the historical patient database, with the physiological indicators of patients as independent variable data and each type of chronic disease risk factor (such as the occurrence of complications) as the dependent variable, a logistic regression model is constructed to obtain the regression coefficients of each indicator for each type of chronic disease risk factor. Then, the absolute values of the regression coefficients are normalized to obtain the contribution ratio of each indicator to each type of chronic disease risk factor. (For example, if the coefficient of a certain indicator for hypertension is 1.5, for diabetes is 0.9, and for chronic kidney disease is 0.6, then the contribution ratio of hypertension is 1.5 / (1.5 + 0.9 + 0.6) = 0.5, for diabetes is 0.3, and for chronic kidney disease is 0.2); (3)Based on the importance of each type of chronic disease risk factor and the contribution ratio of each indicator to each type of chronic disease risk factor, the weighted summation method is used to calculate the real-time risk score of the patient, that is, the importance of each risk factor is multiplied by the corresponding contribution ratio and then added to obtain the risk score. (For example, the risk score of patient A = grade 3 hypertension 0.9×0.5 + diabetes 0.7×0.3 + chronic kidney disease 0.8×0.2 = 0.82); (4)The entropy weight method is applied to dynamically analyze the risk score distribution of all patients. (For example, the median of the current patient group is 0.6), and dynamic thresholds for risk stratification are set according to the risk score distribution (low risk < 0.4, medium risk 0.4 - 0.6, high risk 0.6 - 0.8, very high risk ≥ 0.8); (5)Real-time collect the physiological indicator data of the patient individual (for example, the current systolic blood pressure of patient A is 190 mmHg and the fasting blood glucose is 8.0 mmol / L) and calculate the real-time risk score, and accurately stratify the patient to the corresponding risk level according to the dynamic threshold of risk stratification.

[0030] Table 1 shows the detection frequencies and follow-up items for different cardiovascular disease risk stratifications of hypertensive patients.

[0031] Table 1

[0032] The target population classification and signing management module 11 can perform real-time classification and screening of the target population based on the multi-source and multi-modal database. Through the dynamic risk prediction of the chronic disease prediction model, it will actively identify different types of populations and push relevant management suggestions to chronic disease doctors, helping patients sign up with chronic disease doctors before the onset or in the early stage of the disease and enter the full-process active management stage to better carry out chronic disease prevention and delay the development of the disease. The management path will be adjusted according to the risk level of the patient to ensure that each patient can obtain a personalized medical plan according to their health status.

[0033] In the embodiment, the hierarchical referral path setting module 12 is used to establish the upstream and downstream referral paths, determine the corresponding upstream and downstream medical communities and specific docking medical teams, and specifically provide the hierarchical referral path setting function. Based on the referral path setting function, the upstream and downstream referral paths composed of medical institutions are set. The set upstream and downstream referral paths include: community health service stations (pre-screening) at level 1 → community health service centers (basic diagnosis and treatment) at level 2 → district-owned hospitals (specialist diagnosis and treatment) at level 3 → tertiary hospitals (complex diseases) at level 4. It should be noted that critically ill patients can skip some levels. Through the hierarchical referral path setting function, the responsible doctors corresponding to chronic diseases are also bound to the medical institutions at each level to form a referral relationship mapping table.

[0034] Specifically, in the referral path setting, when the super administrator logs in, the referral path needs to be set, that is, the referral chain of each level of hospital, to clarify the upstream and downstream referral hospitals and the responsible doctors, and a referral hospital is also specifically implemented for high-risk chronic disease patients; when the hospital administrator logs in, the list of medical staff teams for each chronic disease needs to be filled in to correspond to the hospital referral chain.

[0035] In the pre-diagnosis stage of the patient, with the help of nurses, attending doctors, chronic disease doctors, and contracted doctors, the referral chain of the hospital corresponding to the contracted doctor and this patient is confirmed (a certain referral hospital corresponding to the three levels of village health centers, township health centers, and county-level hospitals).

[0036] In the embodiment, the active referral module 13 is used to generate referral suggestions based on the dynamic risk changes of chronic diseases for each type of population and push them to the chronic disease doctors. The chronic disease doctors determine the referral suggestions and set the referral information based on the upstream and downstream referral paths and the referral relationship mapping table, and synchronously notify the target medical institutions and patients of the referral.

[0037] Specifically, in the risk prediction, when the diagnosed population or the population to be managed is determined to be continuously at high risk or the follow-up is not completed based on the dynamic risk changes of chronic diseases, it is determined to be a high-risk situation. Continuously high-risk high-risk populations will also be judged as high-risk situations.

[0038] Among them, continuously being at high risk is evaluated as being continuously identified as having a high risk of disease occurrence prediction or a high risk of disease progression prediction multiple times. The criteria for disease occurrence are formulated by a clinical multi-disciplinary doctor team in combination with guidelines and AI logic. The evaluation criteria for disease progression include disease grading progression, target organ damage, complication occurrence, hospitalization, death, etc. The specific number of consecutive incomplete follow-ups is determined by the severity of the disease.

[0039] For patients determined to be in high-risk situations, referral suggestions will be automatically generated. Specifically, a referral suggestion generation model is used to generate referral suggestions for people in high-risk situations and push them to chronic disease doctors. The referral suggestions include a patient risk analysis report containing an index trend chart and risk assessment, a recommended referral hospital and doctor matched according to a pre-set hierarchical referral path, a real-time emergency score for the patient, an overview of the medical resource status of the referral hospital, and a recommended referral time dynamically recommended based on the source data of the superior hospital's appointment slots.

[0040] The process of generating referral suggestions by the referral suggestion generation model is as follows: (1) Calculate the medical gap index based on the level difference and disease treatment capacity between the patient's current medical institution and the target medical institution. The medical gap index is the product of the level gap between the two medical institutions and the disease treatment capacity. For example, if the level difference between the current secondary hospital and the target tertiary hospital is 1 level, the medical gap index is 1×1.5 = 3, where 1.5 is the disease treatment capacity of the given high-level medical institution. (2) Real-time monitor the patient's key physiological indicators and complication risk scores, and input the monitoring data combined with the medical gap index into a gated recurrent neural network (GRU) to output the emergency score in real time (range 0 - 100 points, and a patient with a score exceeding 80 points is defined as a high-risk patient). (3) Input the emergency score into an asymmetric priority algorithm. For the emergency score of high-risk patients, for every increase in the given score, the queuing priority weight will increase by a given multiple correspondingly. For example, for high-risk patients with an emergency score ≥80, for every increase of 5 points, the queuing priority weight will exponentially increase by 1.5 times, thereby updating the priority order of the patient queue. (4) Collect the medical resource status of the target hospital (such as outpatient appointment rate 85%, registration rate 80%, doctor scheduling, etc.) to generate a resource heat map. Based on the patient sequence with updated priorities, input the patient's priority and the hospital resource heat map into a multi-objective integer programming model. The multi-objective integer programming model takes the patient's priority as the input variable and the medical resource status of the target hospital as the constraint condition to construct an objective function. The objective function aims to maximize the sum of the products of the patient's priority weight (representing the patient's emergency or demand level, such as high-risk 0.9, ordinary 0.5) and the medical resource satisfaction (calculated based on the hospital resource heat map, such as the comprehensive score of outpatient appointment situation, doctor scheduling, etc., for example, 0.9 when resources are abundant, 0.3 when resources are tense). Then, use integer programming (the variable takes values of 0 or 1, representing whether to select this referral path) to solve the model to obtain the optimal referral path, achieve the best match between patient needs and medical resources, and determine the optimal referral path between patient needs and medical resources (such as from a secondary hospital → tertiary hospital A → tertiary hospital B). (5)Automatically generate a complete referral report, including the optimal referral path, the patient's real-time emergency score (such as 87 points, high risk), an overview of the medical resource status of the referral hospital (the outpatient appointment rate of tertiary hospital A is 85%, and the registration rate is 80%), the expected reduction in risks after referral (such as a 25% decrease in the risk of complications), and the reasons for the referral recommendation.

[0041] Among them, the criteria for upward referral of chronic diseases based on which the referral recommendation generation model generates referral recommendations include: ① The system first discovers that a high-risk population has actually fallen ill but has not been diagnosed by a doctor; ② The confirmed population or the population to be managed has a high risk of continuous disease progression; ③ The confirmed population or the population to be managed has a high risk of continuous disease progression; ④ The confirmed population or the population to be managed has not reached the disease control standard continuously; ⑤ The chronic diseases of the confirmed population or the population to be managed have an acute attack or critical illness.

[0042] After receiving the referral recommendation and viewing the prediction basis, the chronic disease doctor selects to accept the referral recommendation or make manual modifications to confirm and generate the final referral information, that is, set the information according to the upstream and downstream referral paths and the referral relationship mapping table to generate the final referral information. Among them, manual modification takes precedence over the generated referral recommendation and has the final decision-making power; when the chronic disease doctor refuses the referral recommendation, the reason also needs to be filled in manually and an electronic signature is required.

[0043] The active referral module 13 automatically sends a referral notice to the patient and the superior doctor, providing the information of the receiving doctor. The responsible doctor needs to guide the patient to seek medical treatment and record the completion of the referral. After the superior doctor receives the patient, the system records the treatment plan and feeds it back to the original responsible doctor. That is, real-time recording is also carried out when receiving the referral and generating the treatment plan at the target medical institution, and the treatment plan is synchronized to the chronic disease doctor.

[0044] In the artificial intelligence-based hierarchical and classified chronic disease management and referral system of the present invention, the active referral process assisted by artificial intelligence has characteristics such as prevention, high efficiency, and precision, improving the response speed and accuracy of medical services.

[0045] In the embodiment, the reverse incentive module 14 is used to perform performance incentives on whether the chronic disease doctor completes the signing of chronic disease patients, implements hierarchical referral, and saves medical expenses. Through the reverse incentive mechanism, the initiative and enthusiasm of the medical staff team are stimulated, and the referral path and management are continuously optimized.

[0046] The key points for evaluating chronic disease doctors lie in whether they have completed the signing of chronic disease patients, implemented hierarchical referral, and saved medical expenses. The system independently records the specific situation of referrals, the results of model predictions, whether referrals are made and modified records, the completion of follow-up visits, and the effectiveness of referrals. Based on the data recorded by the system, evaluations are conducted according to the allocated performance. If chronic disease doctors save medical expenses, they will receive positive performance incentives according to the proportion of medical expenses. On the contrary, they will be deducted to a certain extent. Specifically, the referral acceptance rate and the amount of medical expense savings are statistically calculated, and the performance incentive amount is calculated by combining the basic salary, signing rate, referral acceptance rate, and the amount of medical expense savings.

[0047] Among them, the amount of medical expense savings = the average expense of patients of the same type in the same period last year - the current actual expense; the referral acceptance rate = the number of recommended referrals / the number of referrals actually made by the doctor.

[0048] The performance incentive amount = basic salary × (1 + signing rate × a) + (referral acceptance rate × b%) + (the amount of medical expense savings × c yuan) + d, where a, b, c, and d are all preset coefficients or amounts, and are adjusted regionally according to the regional finance and health service status.

[0049] The hierarchical chronic disease management and referral system based on artificial intelligence provided by the above embodiments, through artificial intelligence-assisted hierarchical referral, predicts according to the disease progression, assists the intelligent flow of patients up and down the referral system, optimizes the allocation and utilization of medical resources, constructs a compact medical community, and realizes the complementary advantages of urban and rural medical care; screens the target population for chronic disease management, implements classified management, reduces the incidence and fatality and disability rates of chronic diseases, and alleviates the heavy medical burden caused by chronic diseases.

[0050] The hierarchical chronic disease management and referral system based on artificial intelligence provided by the above embodiments optimizes the process and monitoring of hierarchical diagnosis and treatment of chronic diseases, reduces the disease burden caused by chronic diseases due to fatality and disability, reduces personal hygiene expenditures and the burden of medical costs; transforms the throttling of medical expenses into performance feedback, and stimulates the initiative of the medical staff team.

[0051] The embodiment also provides a chronic disease management and referral method applying the above hierarchical chronic disease management and referral system, such as Figure 2As shown, according to the clinical guidelines related to chronic diseases, algorithms are used to screen out chronic management targets. Among them, the number of patients with diagnosed chronic diseases is determined based on doctors' diagnoses. The number of patients awaiting chronic disease management is screened from the suspected patient population that has not been effectively included in management. The high-risk population for chronic diseases is a high-incidence risk group determined based on clinical results after risk assessment. For patients with diagnosed chronic diseases and those awaiting chronic disease management, ensure that they are signed and bound to the chronic disease doctor management system. For the high-risk population for chronic diseases, adopt a strategy mainly based on supervision and education to achieve signing and binding as much as possible, forming an active management mode of artificial intelligence. On the basis of the traditional manual referral mode from the medical insurance system, the community service family doctor system to the district-owned hospitals, introduce an active referral mechanism assisted by artificial intelligence. When the patient's condition changes are monitored, quickly start the condition assessment program, make a preliminary judgment on whether to refer, and submit the result to the doctor for review. If the judgment of artificial intelligence is inconsistent with that of the doctor, execute the doctor's judgment and confirm it with an artificial signature. Finally, promptly refer the continuously high-risk or non-followed-up population to the superior hospital.

[0052] Taking the medical communities in a certain district as an example, 1.26 million medical treatment data of the district are included (including medical institution data, public health data, physical examination data, and medical insurance data). Through the target population classification and signing management module, a chronic disease model is used for dynamic risk prediction, and 400,000 chronic disease management target numbers are screened out. Among them, there are 180,000 diagnosed patients who have been diagnosed with chronic diseases by doctors, 60,000 patients awaiting management who actually meet the chronic disease diagnosis criteria but have not been diagnosed by doctors, and 160,000 high-risk patients who do not currently meet the chronic disease diagnosis criteria but are identified as high-risk by AI. In the hierarchical referral path setting module, the hospitals are graded and assigned, and the upstream and downstream referral chains are constructed. A separate referral chain is established for high-risk patients, including information such as the responsible doctor, affiliated hospital, and superior referral.

[0053] For chronic diseases such as diabetes, hypertension, chronic kidney disease, and cardiovascular and cerebrovascular diseases, an active management and active referral mechanism is carried out based on the active referral module. Newly managed patients are dynamically included from the medical insurance and family doctor populations. If the patient is continuously high-risk or not followed up, the system will actively refer the patient to the superior district-owned hospital.

[0054] The specific implementation manners described above have elaborated on the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, supplements, equivalent replacements, etc. made within the principle scope of the present invention should be included within the protection scope of the present invention.

Claims

1. A hierarchical chronic disease management and referral system based on artificial intelligence, characterized by: include: The target population classification and contract management module is used to predict the dynamic risk of chronic diseases based on the multi-source and multi-modal data of the target population and classify the target population, generate chronic disease management suggestions for each group of people and push them to chronic disease doctors for contract management; A hierarchical referral path setting module is used to provide a hierarchical referral path setting function, based on which an upstream and downstream referral path consisting of medical institutions is set, and the responsible doctors corresponding to chronic diseases are bound to each level of medical institutions to form a referral relationship mapping table; Active referral module, which is used to generate referral suggestions based on the dynamic risk changes of chronic diseases in each group of people and push them to chronic disease doctors. Chronic disease doctors determine referral suggestions and set referral information based on upstream and downstream referral paths and referral relationship mapping tables, and simultaneously notify the target diagnosis and treatment institutions and patients for referral; The reverse incentive module is used to provide performance incentives to chronic disease doctors based on whether they have completed the signing of chronic disease patients, implemented tiered referrals, and saved medical expenses.

2. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: In the target population classification and contract management module, the dynamic risk of chronic diseases is predicted based on the multi-source and multi-modal data of the target population, and the target population is classified. Chronic disease management suggestions are generated for each type of population and pushed to chronic disease doctors for contract management, including: Use chronic disease prediction models to predict the dynamic risk of chronic diseases based on multi-source and multi-modal data of the target population; The target population whose dynamic risk prediction results of chronic diseases are lower than the first threshold are classified as healthy people. For the healthy population, follow-up reminders and follow-up items are regularly pushed to chronic disease doctors and patients, and the follow-up item data of the healthy population are actively tracked, and the prediction of progression to high-risk population is observed; The target population whose chronic disease dynamic risk prediction results exceed the second threshold is classified as a high-risk population, and chronic disease doctors are signed for the high-risk population on the principle of signing as many as possible. The follow-up frequency is increased according to the personalized disease risk of high-risk individuals, and the follow-up frequency data of the high-risk population is actively tracked. If the follow-up is not completed within a certain number of expected follow-up intervals, a warning value will be triggered and chronic disease doctors will be actively reminded to pay attention to high-risk individuals; The population with clinically confirmed diseases is classified as the confirmed population, and the population who are actually ill but have not been diagnosed by a doctor is classified as the population to be managed. Both the confirmed population and the population to be managed are required to sign contracts with chronic disease doctors, and sign contracts and long-term management are achieved through telephone mobilization, APP reminders, tracking of medical records, and high-risk health education.

3. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 2 is characterized in that: The chronic disease prediction model is used to predict the dynamic risk of chronic diseases based on multi-source and multi-modal data of the target population, including: First, based on three types of structured multimodal data of the target population, namely historical diagnostic data, biochemical indicators and follow-up information, a multimodal health trajectory in the form of a time series is constructed; Secondly, the chronic disease prediction model adopts a fusion method of dynamic risk encoding mechanism and attention mechanism. The multimodal health trajectory data is input into the chronic disease prediction model according to a certain time window. It is first encoded and the attention mechanism is used on the encoding features to capture the nonlinear change characteristics of the key variables of interest in the data under different time windows. The convolutional neural network and long short-term memory network used in the multi-task learning head are used to fuse the nonlinear change characteristics of each key variable, and the dynamic risk of chronic diseases and health status trends in the future time period are predicted based on the fusion results. The dynamic risk of chronic diseases is expressed in the form of probability, and the larger the probability, the higher the risk. An adaptive time decay function is also embedded in the chronic disease prediction model, which introduces a weight reduction mechanism for distant historical data. Different decay weights are set for the input data in the time window according to the distance from the prediction time to adjust the degree of influence on the prediction results. Specifically, the farther the input data is from the prediction time, the larger the decay weight is set.

4. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 2 is characterized in that: Long-term management of confirmed cases and those awaiting care includes: Carry out graded management according to the progress of chronic diseases, specifically based on the consensus and guidelines of the clinical multidisciplinary doctor team as prior knowledge. For patients with different chronic disease progression, the cardiovascular disease risk stratification is determined according to the cardiovascular disease risk stratification model, which includes low risk, medium risk, high risk, and very high risk. Follow-up frequencies and follow-up items with different requirements are carried out according to the stratification results. Among them, the risk level of disease progression is determined according to the risk stratification model, including: (1) Construct a knowledge map based on the consensus of clinicians to clarify the risk factors of chronic diseases and their importance; (2) Using the historical patient database, with the patient's physiological indicators as the independent variable data and each type of chronic disease risk factor judged by complications as the dependent variable, a logistic regression model was constructed to obtain the regression coefficient of each indicator for each type of chronic disease risk factor. The absolute value of each regression coefficient was then normalized to obtain the contribution ratio of each indicator to each type of chronic disease risk factor. (3) Based on the importance of each type of chronic disease risk factor and the contribution of each indicator to each type of chronic disease risk factor, the patient's real-time risk score is calculated using the weighted summation method; (4) Apply the entropy weight method to dynamically analyze the risk score distribution of all patients and set the dynamic threshold of risk stratification based on the risk score distribution; (5) Collect individual patient physiological indicator data in real time and calculate real-time risk scores, and accurately stratify patients into corresponding risk levels according to the dynamic thresholds of risk stratification.

5. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: The hierarchical referral path setting module is used to set the upstream and downstream referral paths when the super administrator with high authority logs in, and the set upstream and downstream referral paths include: community health service stations at level 1 → community health service centers at level 2 → district hospitals at level 3 → tertiary hospitals at level 4; referral hospitals and target responsible doctor teams are also implemented separately for high-risk groups of chronic diseases; When the chronic disease doctor logs in, the hierarchical referral path setting module fills in the list of superior referral doctors, and the nurse fills in the assistance management path.

6. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: In the active referral module, referral suggestions are generated based on the dynamic risk changes of chronic diseases in each group of people and pushed to chronic disease doctors, including: When the confirmed population or the population to be managed is judged to be at continuous high risk based on the dynamic risk changes of chronic diseases, or the follow-up is not completed, it is judged to be a high-risk situation. When the high-risk population is continuously at high risk, it will also be judged to be a high-risk situation; The referral recommendation generation model is used to generate referral recommendations for people in high-risk situations and push them to chronic disease doctors. The referral recommendations include patient risk analysis reports containing indicator trend charts and risk assessments, recommended referral hospitals and doctors matched according to pre-set hierarchical referral paths, and recommended referral times dynamically recommended based on the number source data of higher-level hospitals.

7. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 6 is characterized in that: The process of generating referral recommendations for people at high risk is as follows: (1) Calculate the medical gap index based on the difference in grades and disease handling capabilities between the patient's current medical institution and the target medical institution, where the medical gap index is the product of the grade difference between the two medical institutions and the disease handling capabilities; (2) Monitor the patient's key physiological indicators and complication risk scores in real time, and input the monitoring data into the gated recurrent neural network in combination with the medical gap index, output the emergency score in real time, and identify high-risk patients based on the emergency score; (3) Input the emergency score into the asymmetric priority algorithm. For high-risk patients, the priority weight of the queue is increased by a given multiple for each increase in the emergency score, thereby updating the priority order of the patient queue. (4) The medical resource status of the target hospital is collected and a resource heat map is generated. Based on the patient sequence after the priority is updated, the patient priority and the hospital resource heat map are input into a multi-objective integer programming model. The multi-objective integer programming model uses the patient priority as the input variable and the medical resource status of the target hospital as the constraint condition to construct an objective function. The objective function aims to maximize the sum of the product of the patient's priority weight and the medical resource satisfaction. Then, integer programming is used to solve the model to obtain the optimal referral path, achieve the best match between patient needs and medical resources, and determine the optimal referral path between patient needs and medical resources. (5) Automatically generate referral reports, including the optimal referral path, the patient's real-time urgency score, an overview of the medical resource status of the referral hospital, the expected risk reduction after referral, and the reasons for the referral recommendation.

8. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: In the active referral module, the chronic disease doctor determines the referral suggestion and sets the referral information based on the upstream and downstream referral paths and referral relationship mapping table and simultaneously notifies the target diagnosis and treatment institution and patient of the referral, including: After receiving the referral suggestion and checking the prediction basis, the chronic disease doctor chooses to accept the referral suggestion or manually modify it to confirm and generate the final referral information, that is, to set the information according to the upstream and downstream referral paths and referral relationship mapping table to generate the final referral information, among which manual modification takes precedence over the generated referral suggestion; When chronic disease doctors reject referral recommendations, they must manually fill in the reasons and sign electronically.

9. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: The active referral module also keeps real-time records when the target diagnosis and treatment institution accepts the referral and generates a treatment plan, and synchronizes the treatment plan to the chronic disease doctor.

10. The artificial intelligence-based hierarchical chronic disease management and referral system according to claim 1 is characterized in that: In the reverse incentive module, performance incentives are provided to chronic disease doctors based on whether they have completed the signing of chronic disease patients, implemented graded referrals, and saved medical expenses, including: The referral acceptance rate and medical cost savings are calculated based on the specific referral circumstances recorded in the system, the model prediction results, whether the referral is made and the record is modified, the follow-up completion status, and the referral effectiveness. The performance incentive amount is calculated based on the basic salary, contract signing rate, referral acceptance rate, and medical cost savings.

Images