WO2023109250A1

WO2023109250A1 - Internet-based hospital triage data processing method and system

Info

Publication number: WO2023109250A1
Application number: PCT/CN2022/121815
Authority: WO
Inventors: 张颖聪; 武青松; 马鸣; 向璨; 陈实; 吴建才; 金阳; 王征; 罗飞; 王智慧
Original assignee: 华中科技大学同济医学院附属协和医院
Priority date: 2021-12-15
Filing date: 2022-09-27
Publication date: 2023-06-22
Also published as: CN114171176A

Abstract

The present disclosure belongs to the technical field of the Internet. Disclosed are an Internet-based hospital triage data processing method and system. The system comprises an appointment management platform, wherein an output end of the appointment management platform is connected to an input end of a triage registration unit. Identity information, physical condition information and case information of a patient are collected by means of an appointment management platform, and an illness state keyword is extracted from physical condition information data of the patient on the basis of a case evaluation model, so as to evaluate a symptom; moreover, information sharing among different departments and different doctors is realized, a triage scheme is intelligently recommended, and association matching is effectively performed on an expert and the patient according to the specific case information of the patient, thereby improving the triage efficiency and accuracy while reducing the workload.

Description

An Internet-based hospital triage data processing method and system

Cross References to Related Applications

This disclosure claims the priority of a Chinese patent application filed on December 15, 2021 with application number 202111533487.6 and titled "An internet-based hospital triage data processing method and system", the entire contents of which are hereby incorporated by reference .

technical field

The disclosure belongs to the technical field of the Internet, and specifically relates to an Internet-based hospital triage data processing method and system.

Background technique

The public triage refers to the rapid and focused collection of data on patients who come to the emergency department of the hospital, and the data are analyzed, judged, classified, and divided into departments, and at the same time, the sequence of visits is arranged according to light, heavy, slow, and urgent, and registered in the register (file) ), the time should generally be completed within 2 to 5 minutes. The focus of triage: disease triage and subject triage.

Obtain the subjective data of the patient through interrogation, that is, the chief complaint and related accompanying symptoms, and understand the patient's feelings about the disease, psychological state, behavioral response, and social situation, and understand the past history, medication history, and allergy history related to the current medical history. etc. During the consultation process, we should pay attention to the tendentious statements of patients and their family members, and conduct purposeful interviews according to the condition, so that the collected information is true and comprehensive. Interrogate or imply the patient, so as not to make the information of the consultation inconsistent with the actual situation, and do not bring adverse stimulation or adverse influence to the patient's spirit. For example, patients with acute abdominal pain should pay attention to asking whether they have diarrhea, overeating or eating spicy food. Women especially ask about their menstrual status, fearing the occurrence of ectopic pregnancy. With the rapid development of Internet and artificial intelligence technology, text recognition and classification are applied in more and more fields. For example, the triage model used to provide intelligent guidance services for hospitals, its input is the patient's condition description, and the output is the department number that the patient should register, which is a typical text classification problem. Due to the particularity of the triage business of the hospital's departments, patients cannot be misled to go to the wrong department, so the accuracy of the triage model is very high.

At present, the process of seeing a doctor in a hospital can be done through on-site registration, online appointment registration or telephone appointment registration. When making an appointment by telephone or on-site registration, the registrar will perform triage according to the patient's simple self-report and select different departments. During online registration, patients choose departments and doctors according to their own experience during the registration process.

At present, during the registration process, experts and doctors can make appointments according to different departments. Since the experts and doctors cannot fully understand the patient's case during the appointment process, the patient can only choose the designated one according to the case. For experts and physicians, some patients may directly book an expert number for triage due to lack of medical knowledge. In the process of triage, the pressure on experts will be increased, and high-quality medical resources will be wasted. Correlation matching leads to inaccurate triage results and affects subsequent diagnosis and treatment.

During the appointment and triage process of patients, the locations of each department are relatively widely distributed, and each department has access control checkpoints. Due to the limited appointment time, it is impossible to judge whether the patient has arrived at the designated department for treatment and remind in time according to the patient's actions. Patients, and the various procedures in the hospital are complicated, and the places for the procedures are relatively scattered. When arriving at the designated place, it is easy to cause a certain procedure to be turned back if it is not completed, which will cause certain troubles to the patients and delay the time for seeing a doctor.

Contents of the invention

In order to overcome the above-mentioned defects of the prior art, the present disclosure provides a method and system for processing hospital triage data based on the Internet. By using one or more implementations of the present disclosure, it solves the problem that experts and physicians cannot fully understand patients. According to the specific case conditions, patients can only choose designated experts and doctors according to the case conditions. Due to the lack of medical knowledge, some patients will directly make appointments with expert numbers for triage. In the process of triage, the work pressure of experts will increase, and high-quality medical resources will appear It is wasteful. It is impossible to associate and match experts and patients based on the patient's specific case information, and it is impossible to judge whether the patient has arrived at the designated department for medical treatment based on the patient's actions and to remind the patient in time.

In order to achieve the above purpose, according to some embodiments of the present disclosure, an Internet-based hospital triage data processing method is provided, including the following steps: S1, first collect patient identity information, physical condition information, and case information; S2, establish a case Evaluate the model, and bring the collected data information into the case evaluation model; S3, generate four evaluation indicators of mild condition A, acute condition B, severe condition C and critical condition D through the case assessment model; S4, establish a triage guide Diagnosis model, responsible for the patient triage process, based on creating tasks on the appointment management platform and assigning tasks; S5. Send the obtained triage plan to the patient receiving data terminal through the information publishing module, and let the patient judge whether to implement the triage plan ; S6. During the triage process, identify the face feature information through the department access control identification terminal, perform patient identity verification, and generate a location state chain.

According to some embodiments of the present disclosure, an Internet-based hospital triage data processing system is provided, including an appointment management platform, the output end of the appointment management platform is connected to the input end of the triage registration unit, and the triage registration unit The output end of the unit is connected to the input end of the identity authentication module, the input end of the identity authentication module is connected to the input end of the waiting queue management module, the input end of the waiting queue management module is connected to the input end of the reservation management platform, and the The appointment management platform includes a case evaluation model, a patient information database, a medical personnel database, and a case medical database, and the triage registration unit includes a triage guidance model and an information release module, and the input terminal of the information release module is connected to the triage guide. Connect to the output terminal of the diagnostic model.

In some embodiments of the present disclosure: the appointment management platform is used to collect patient identity information, physical condition information, and case information, and the case evaluation module is used to extract disease information keywords from the patient's physical condition information. Word formation evaluation results, the case evaluation model is divided into four evaluation indicators of mild condition, acute condition, severe condition and critical condition, respectively A, B, C, D, reflecting the severity of the condition of each patient.

In some embodiments of the present disclosure: the case evaluation model includes a case information module, a case search module, a keyword recognition module, and an index information entry module. The steps for establishing the case evaluation model are as follows: S101. First, through the case information module Input personal basic information, including consultation record information, case report information, medication and adjustment information, and automatically generate the basic information number of each patient, enter and authenticate the keywords contained in each information through the keyword identification module, and pass the keyword The search information input by module recognition promptly displays all information about keywords; S102, secondly, find all the information of patients with corresponding numbers through the case search module for the patient's name, number, gender, age and date of birth; S103, according to the search condition The keyword data information associated with mild, urgent, severe and critical conditions can list A, B, C, and D evaluation index lists, and obtain evaluation results for personnel to arrange medical treatment in time.

In some embodiments of the present disclosure: the evaluation method in S103 is as follows: S1031, compare the disease description sentence with each disease description information in the preset disease keywords, and determine the disease description sentence key that matches the sentence words, and similar disease description keywords correspond to A, B, C, and D disease evaluation indicators; S1032, input the determined disease description keywords into the case evaluation model, and the case evaluation model outputs the triage departments.

In some embodiments of the present disclosure: the triage guidance model includes a data fusion module, a triage training module and a triage scheduling module, and the establishment method of the triage guidance model is as follows: E1, the condition information of the patient and the disease level are collected, and the modified cosine method is used to intelligently fuse the data of the disease status, and the fused data is transmitted to the triage training module. The calculation formula is:

And y and z represent the data of the pathological condition, represent the score of m on all the collected data, M _my and M _mz represent the score of m on data y and data z respectively, K _yz represents the score of data y and data z Score similarity; E2. The training module trains the disease data and transfers the data to the triage scheduling module. The terminals of each triage department connected to the triage scheduling module store the pathological data of patients received and fused in a structural format. The structural format includes 【G(t), g(t), G(t+1)】, which respectively represent the state G of the current stage t of the condition, the triage plan g executed in the current stage, and the condition of the next visit stage t+1 Information; if the similarity with the state G of the current stage t of the disease state is greater than the threshold of the pathological state, then output the corresponding triage plan g; E3, send the triage plan obtained in E2 to the patient through the information release module to receive data On the terminal, the patient can judge whether to implement the triage plan, if so, execute the condition and send the confirmation information to the appointment management platform, if not, re-enter the condition information or communicate with relevant personnel in time.

In some embodiments of the present disclosure: the training module is established in the following manner: E201, for multiple disease description texts of each department, carry out keyword numbering on a single disease description text, and obtain the primary disease state of each disease description text The feature vector is processed for the condition keyword table of each department to obtain the keyword table vector for each department; E202, for each condition description text, compare the feature vector of the condition description text with the keyword table vector of each department The correlation between generates the secondary feature vector of the description text of the condition for each department;

E203. For each disease description text, compare the correlation between the primary feature vector and the secondary feature vector of the disease description text, generate a fusion feature vector of the disease description text, and match the fusion feature vector and the disease description text of each disease description text department information, to obtain the triage training module; in the E201, the correlation formula between the disease description text s and the department explanation text is:

PMI(s) represents the correlation between the disease description text s and the departmental explanation text, and F(s, m) represents the joint distribution probability of the description text s and the disease feature vector m, which can characterize the frequency of both occurrences at the same time.

In some embodiments of the present disclosure: the identity authentication module is used to identify the patient's identification card information or face feature information, match the corresponding patient ID according to the identified information, obtain the patient's location information, and obtain the patient's location information according to the multiple location information Determines the location state chain that generates the patient.

In some embodiments of the present disclosure: the waiting queue management module obtains the queuing information of each department or doctor in the clinic area from the HIS system, manually increases the patient queue according to the registration information, and generates the general outpatient waiting queue and Specialist outpatient waiting queue.

Description of drawings

Fig. 1 shows a schematic block diagram of an Internet-based hospital triage data processing system according to some embodiments of the present disclosure;

Fig. 2 shows a simplified flowchart of an Internet-based method for processing hospital triage data according to some embodiments of the present disclosure.

Detailed ways

The technical solution of this patent will be further described in detail below in conjunction with specific embodiments.

As shown in Figures 1-2, in some embodiments of the present disclosure, an Internet-based hospital triage data processing method is provided, and the hospital triage data processing method may include the following steps: S1, first collect patient identity information , physical condition information, case information. S2. Establish a case evaluation model, and bring the collected data information into the case evaluation model. S3. Generate four evaluation indicators of mild condition A, acute condition B, severe condition C and critical condition D through the case assessment model. S4. Establish a triage and guidance model, be responsible for the patient triage process, create and assign tasks based on the appointment management platform. S5. Send the obtained triage plan to the patient receiving data terminal through the information publishing module, and let the patient judge whether to implement the triage plan. S6. During the triage process, identify the facial feature information through the department access control identification terminal, perform patient identity verification, and generate a location status chain.

In some embodiments of the present disclosure, an Internet-based hospital triage data processing system is provided, and the hospital triage data processing system may include an appointment management platform, an output terminal of the appointment management platform and an input of the triage registration unit The output end of the triage registration unit is connected to the input end of the identity authentication module, the input end of the identity authentication module is connected to the input end of the waiting queue management module, and the input end of the waiting queue management module is connected to the input end of the appointment management platform , the reservation management platform includes a case evaluation model, a patient information database, a medical staff database, and a case medical database. The triage registration unit includes a triage guidance model and an information release module, and the input terminal of the information release module is connected to the output of the triage guidance model. end connection.

In some embodiments of the present disclosure, the appointment management platform can be used to collect patient identity information, physical condition information, and case information, and the case evaluation module is used to extract disease information keywords from the patient's physical condition information, and form an evaluation based on the disease keywords As a result, the case evaluation model is divided into four evaluation indicators: mild condition, acute condition, severe condition and critical condition, respectively A, B, C and D, reflecting the severity of each patient's condition.

In some embodiments of the present disclosure, the case evaluation model may include a case information module, a case search module, a keyword identification module, and an index information entry module. The steps for establishing the case evaluation model may be as follows: S101. Basic information, including consultation record information, case report information, medication and adjustment information, and automatically generate the basic information number of each patient, enter and authenticate the keywords contained in each information through the keyword identification module, and identify them through the keyword module The search information entered will display all the information about the keyword. S102. Next, through the case search module, find all the information of the patient with the corresponding number on the patient's name, number, gender, age and date of birth. S103 , according to the keyword data information related to the search for mild condition, urgent condition, severe condition and critical condition, the list of evaluation indicators A, B, C, and D can be listed, and the evaluation results can be obtained for personnel to arrange medical treatment in time.

In some embodiments of the present disclosure, the evaluation method in S103 can be as follows: S1031, compare the condition description sentence with each condition description information in the preset condition keywords, and determine the condition description sentence keywords that match the statement , and similar disease description keywords correspond to A, B, C, and D disease evaluation indicators respectively. S1032. Input the determined disease description keywords into the case evaluation model, and the case evaluation model outputs the triage departments.

In some embodiments of the present disclosure, the triage guidance model may include a data fusion module, a triage training module, and a triage scheduling module, and the establishment method of the triage guidance model may be as follows: The level is collected, and the modified cosine method is used to intelligently fuse the data of the patient's condition, and the fused data is transmitted to the triage training module. The calculation formula is:

And y and z represent the data of the pathological condition, represent the score of m on all the collected data, M _my and M _mz represent the score of m on data y and data z respectively, K _yz represents the score of data y and data z Score similarity. E2. The training module trains the disease data and transmits the data to the triage scheduling module. The triage department terminals connected to the triage scheduling module store the pathological data of patients received and fused in a structural format, which includes [G(t ), g(t), G(t+1)】, respectively represent the state G of the current stage t of the condition, the triage plan g executed in the current stage, and the condition information of the next visit stage t+1. If the similarity with the state G of the current stage t of the disease state is greater than the threshold of the pathological state, then output the corresponding triage plan g. E3. By sending the triage plan obtained in E2 to the patient receiving data terminal through the information release module, the patient can judge whether to implement the triage plan, if so, execute the condition and send the confirmation information to the appointment management platform, if not, re-enter the condition Information or timely communication with relevant personnel.

In some embodiments of the present disclosure, the training module can be established in the following manner: E201. For multiple disease description texts of each department, perform keyword numbering on a single disease description text, and obtain the primary characteristics of the disease in each disease description text The vector is processed according to the disease keyword table of each department, and the keyword table vector for each department is obtained. E202. For each condition description text, compare the correlation between the feature vector of the condition description text and the keyword table vector of each department, and generate a secondary feature vector of the condition description text for each department. E203. For each disease description text, compare the correlation between the primary feature vector and the secondary feature vector of the disease description text, generate a fusion feature vector of the disease description text, and match the fusion feature vector and the disease description text of each disease description text Department information, get the triage training module.

In some embodiments of the present disclosure, the correlation formula between the disease description text s and department explanation text in E201 may be:

In some embodiments of the present disclosure, the identity authentication module can be used to identify the patient's identification card information or facial feature information, match the corresponding patient ID according to the identified information, and obtain the patient's location information. Generate the patient's position state chain.

In some embodiments of the present disclosure, the waiting queue management module can obtain the queuing information of each department or doctor in the clinic area from the HIS system, manually increase the patient waiting queue according to the registration information, and generate the general outpatient waiting queue and expert doctor according to the department registration information. Outpatient waiting queue.

To sum up: according to some implementations of the present disclosure, the patient’s identity information, physical condition information, and case information can be collected through the appointment management platform, based on the case evaluation model and the disease keywords can be extracted from the patient’s physical condition information data, and the disease can be evaluated , and perform triage registration according to the types of diseases classified by the disease indicators. Through the establishment of the triage guidance model in the triage registration unit, the patient's condition is recorded. The condition is compared with the database and recommended to match with the patient. triage plan, and achieve information sharing between different departments and different doctors, intelligently recommend triage plans, effectively match experts and patients according to specific case information of patients, reduce workload and improve triage efficiency and accuracy .

According to some embodiments of the present disclosure, the identity authentication module can also be used to identify the patient's identification card information or face feature information, and match the corresponding patient ID according to the identified information to obtain the patient's location information. Determine and generate the patient's location state chain, and determine whether the patient has arrived at the corresponding department for treatment according to the patient's location state chain. Further improve the efficiency of medical treatment.

The preferred implementation of this patent has been described in detail above, but this patent is not limited to the above-mentioned implementation. Within the scope of knowledge of those of ordinary skill in the art, various implementations can be made without departing from the purpose of this patent. kind of change.

Claims

A method for processing hospital triage data based on the Internet, comprising the following steps:

S1. First collect patient identity information, physical condition information, and case information;

S2. Establish a case evaluation model, and bring the collected data information into the case evaluation model;

S3. Generate four evaluation indicators of mild condition A, acute condition B, severe condition C and critical condition D through the case evaluation model;

S4. Establish a triage guidance model, be responsible for the patient triage process, create and assign tasks based on the appointment management platform;

S5. Send the obtained triage plan to the patient receiving data terminal through the information publishing module, and the patient judges whether to implement the triage plan;

S6. During the triage process, identify the facial feature information through the department access control identification terminal, perform patient identity verification, and generate a location status chain.
An Internet-based hospital triage data processing system, comprising a reservation management platform, the output end of the reservation management platform is connected to the input end of the triage registration unit, and the output end of the triage registration unit is connected to the input of the identity authentication module The input end of the identity verification module is connected to the input end of the waiting queue management module, and the input end of the waiting queue management module is connected to the input end of the reservation management platform, which includes a case evaluation model, a patient An information database, a medical staff database and a case medical database, the triage registration unit includes a triage guidance model and an information release module, the input end of the information release module is connected to the output end of the triage guidance model.
An Internet-based hospital triage data processing system according to claim 2, wherein the appointment management platform is used to collect patient identity information, physical condition information, and case information, and the case evaluation module is used to The key words of disease information are extracted from the status information, and the evaluation results are formed according to the disease keywords. The case evaluation model is divided into four evaluation indicators: mild disease, urgent disease, severe disease and critical disease, which are A, B, C, and D respectively. , reflecting the severity of each patient's condition.
A kind of internet-based hospital triage data processing system according to claim 3, wherein, said case evaluation model comprises case information module, case search module, keyword identification module, index information entry module, said case evaluation model The establishment steps are as follows:

S101. First, input basic personal information through the case information module, including medical record information, case report information, medication and adjustment information, and automatically generate the basic information number of each patient, and use the keyword identification module to identify keywords contained in each information Perform input authentication, identify the input search information through the keyword module and display all information about the keyword;

S102. Next, find all the information of the patient with the corresponding number on the patient's name, number, gender, age and date of birth through the case search module;

S103 , according to the keyword data information related to the search for mild condition, urgent condition, severe condition and critical condition, the list of evaluation indicators A, B, C, and D can be listed, and the evaluation results can be obtained for personnel to arrange medical treatment in time.
A kind of Internet-based hospital triage data processing system according to claim 4, wherein, the evaluation method in the described S103 is as follows:

S1031. Compare the condition description sentence with each condition description information in the preset condition keywords, determine the condition description sentence keywords matching the statement, and the similar condition description keywords correspond to A, B, C, D respectively Condition evaluation index;

S1032. Input the determined disease description keywords into the case evaluation model, and the case evaluation model outputs the triage departments.
A kind of internet-based hospital triage data processing system according to claim 2, wherein, said triage guidance model comprises a data fusion module, a triage training module and a triage scheduling module, said triage guidance model is created as follows:

E1. Collect the patient's condition information and condition level, use the modified cosine method to intelligently fuse the data of the condition of the patient, and transmit the fused data to the triage training module. The calculation formula is:

And y and z represent the data of the pathological condition, represent the score of m on all the collected data, M my and M mz represent the score of m on data y and data z respectively, K yz represents the score of data y and data z score similarity;

E2. The training module trains the disease data and transmits the data to the triage scheduling module. The triage department terminals connected to the triage scheduling module store the pathological data of patients received and fused in a structural format, which includes [G(t ), g(t), G(t+1)】, respectively represent the state G of the current stage t of the condition, the triage plan g executed in the current stage, and the condition information of the next stage t+1;

If the similarity with the state G of the current stage t of the disease state is greater than the threshold of the pathological state, then output the corresponding triage plan g;

E3. By sending the triage plan obtained in E2 to the patient receiving data terminal through the information release module, the patient can judge whether to implement the triage plan, if so, execute the condition and send the confirmation information to the appointment management platform, if not, re-enter the condition Information or timely communication with relevant personnel.
A kind of internet-based hospital triage data processing system according to claim 6, wherein, said training module is set up in the following way:

E201. For multiple disease description texts of each department, carry out keyword numbering to a single disease description text, obtain the primary feature vector of the disease of each disease description text, process the disease keyword table for each department, and obtain the The keyword table vector of the department;

E202. For each condition description text, compare the correlation between the feature vector of the condition description text and the keyword table vector of each department, and generate a secondary feature vector of the condition description text for each department;

E203. For each disease description text, compare the correlation between the primary feature vector and the secondary feature vector of the disease description text, generate a fusion feature vector of the disease description text, and match the fusion feature vector and the disease description text of each disease description text department information, get the triage training module;

The correlation formula between the disease description text s and the department explanation text in E201 is:

PMI(s) represents the correlation between the disease description text s and the departmental explanation text, and F(s, m) represents the joint distribution probability of the description text s and the disease feature vector m, which can characterize the frequency of both occurrences at the same time.
A kind of Internet-based hospital triage data processing system according to claim 2, wherein, the identity authentication module is used to identify the identification card information or face feature information of the patient, and matches the corresponding patient ID according to the identified information, The patient's location information is obtained, and a patient's location status chain is generated according to the determination of multiple location information.
An Internet-based hospital triage data processing system according to claim 2, wherein the waiting queue management module obtains the queuing information of each department or doctor in the clinic area from the HIS system, and manually adds patients to see a doctor according to the registration information Queues, according to the registration information of departments, generate general outpatient waiting queues and expert outpatient waiting queues.