KR102314332B1 - Medical dialog support system and method for physicians and patient using machine learning and NLP - Google Patents

Abstract

The present invention relates to a system and method for supporting a doctor-patient treatment consultation. The system provides a voice consultation service to a patient or a guardian and at the same time, after extracting medically meaningful information such as a medical history, a symptom progress, and a medication status of the patient from the voice consultation, generates a medical consultation report based on the extracted information, and not only maximizes the efficiency and effectiveness of the treatment consultation when compared with the same treatment time by providing the same to a medical staff and the patient but also accordingly increases the accuracy and reliability of a diagnosis and prescription of a doctor, thereby improving the satisfaction and medication compliance for the treatment consultation of the patient, and reducing the social medical costs for treating the patient with the chronic disease.

Description

Medical dialog support system and method for physicians and patient using machine learning and NLP}

The present invention relates to a medical consultation support system and method for a doctor-patient using machine learning natural language processing. More specifically, the present invention provides a voice counseling service to a patient or guardian and, at the same time, provides a voice counseling service to the patient's medical history, symptom progress, and medication After extracting medically meaningful information such as the situation, a medical consultation report is generated based on the extracted information and provided to the medical staff and patients, thereby maximizing the efficiency and effect of medical consultation compared to the same treatment time and, accordingly, the doctor A system and method for supporting doctor-patient consultation that can increase the accuracy and reliability of diagnosis and prescription, improve patient satisfaction with consultation and medication compliance, and reduce social medical costs for the treatment of chronic disease patients will be.

In recent years, medical facilities such as hospitals and pharmacies and the number of users using medical facilities have increased exponentially as the level of medicine, medical technology and equipment have been developed, and interest in health has rapidly increased. Various studies are being conducted to provide medical services.

However, in contrast to limited medical facilities, as the number of users using medical facilities increases significantly, each medical staff digests a lot of treatment, which shortens the doctor's treatment time per patient, and shortening the treatment time is It not only reduces the accuracy, effectiveness and efficiency of diagnosis and prescription, but also increases patient dissatisfaction with the short treatment time, which is a major cause of lowering the reliability of the overall medical service.

For example, according to a paper published in the Journal of Health Administration in 2017, according to the results of a survey of 6,954 first-time patients who visited the National Health Insurance Ilsan Hospital in August 2016, the average initial consultation time was only 6.2 minutes, and the most In the case of the Department of Psychiatry, which requires a long treatment time, it was 13.9 minutes, but it did not reach 16.2 minutes, which is the treatment time that the actual patient would be satisfied with. The results are described.

In general, in order for a doctor to determine an accurate diagnosis and prescription for a patient's symptoms, it is very important to recognize the meaningful information necessary for the patient's treatment in advance, such as the patient's disease, medical history, symptom course, medication compliance, etc. , thus, doctors dedicate their time for treatment to acquire meaningful information necessary for diagnosis and prescription of the patient through conversation (consultation) with the patient for a predetermined period of time prior to the examination.

In particular, in the case of mental health and diseases, it is very important to inform the doctor of the patient's psychological state and life-related information during consultation. In other words, it is known that the patient's satisfaction with treatment increases as the doctor's conversation time rate for the length of treatment, the patient-doctor's sharing of the patient's disease, and the daily conversation with the patient increase.

However, such medical consultation takes a certain amount of time even if only the questions and answers required between the doctor and the patient are required. In general, in the case of medical consultation, the information required for diagnosis and prescription of the doctor is meaningless ( Due to the time constraint of a doctor who has to perform treatment in a limited time because a large part of unnecessary) dialogue is involved, the doctor may perform diagnosis, diagnosis, and prescription only with essential reconnaissance of the patient's condition.

In particular, in the case of patients with cognitive decline, such as old age, brain surgery, or mental illness, conversations and communication with patients become more difficult due to reasons such as not answering the doctor's questions accurately or having an open conversation with questions. Because it is difficult, the effectiveness of medical consultation is further reduced.

In other words, there is an urgent need for research to increase the accuracy and reliability of diagnosis and prescription by inducing consultation on the contents necessary for diagnosis and prescription during medical consultation between doctors and patients, thereby maximizing the effect of consultation compared to the same time.

Meanwhile, in recent years, various studies have been conducted on chat-bot technology that uses artificial intelligence to perform text conversations and voice conversations with users.

When this chatbot technology is applied to a labor-intensive service, it is widely used in various services due to the advantages of reducing the labor force of the service as well as increasing work efficiency.

However, chatbot technology is still being introduced in a limited way to specific services such as simply providing/guided users with specific information or making reservations/changes/cancellations.

In other words, there is an urgent need for research to increase the effectiveness and efficiency of consultation time of less than 15 minutes by introducing chatbot technology to consultation services for doctors and patients, but there is no research on this yet.

1 is a diagram schematically illustrating a method of providing a treatment reservation consultation chatbot service disclosed in Korean Patent Application Laid-Open No. 10-2020-0141957 (Title of the Invention: Chatbot-based method for providing treatment reservation service).

In the method of providing a consultation consultation chatbot service for treatment reservation (hereinafter referred to as the prior art) of FIG. 1 (S900), when a patient enters the chat room (S901), a message welcoming the entry into the chat room is provided and at the same time, a top-level query is searched (S902) ), configure a menu based on the query word, and then provide it together with a text input window (S903).

Also, when the patient selects a query word based on the menu or inputs a query word through a text input window (S904), lower-level query words subordinate to the corresponding query word are searched (S905). At this time, if lower-level query words subordinate to the corresponding query word are searched, re-enter step 903 (S903) to configure a menu based on the query word, and then repeatedly perform the operation of providing the text input window together (S906).

On the other hand, if guidance information is linked to the corresponding query instead of the lower-level query (S907), the guidance information linked to the corresponding query is read and guided to the patient (S908), and if the task processor is linked (S909), the corresponding task After performing the processor, the work result is notified to the patient (S910).

In addition, although steps 907 ( S907 ) to 910 ( S910 ) are not shown in the drawing, if the query word is a new reservation, asking for a reservation standard, and when a clinic is selected as the reservation standard, a list of clinics is provided, but the reservation standard When a disease is selected by , the step of guiding the list of related departments while guiding the description information of the selected disease; When the desired medical staff and time are determined based on the schedule, a reservation process is performed based on the patient reservation, and then the reservation result is guided to the patient.

The prior art (S900) configured as described above uses a chatbot to not only perform patient consultation without a separate intervention by a counselor, but also implements and provides a chatbot system that supports both the natural language processing method and the rule-based menu method. And work efficiency is improved, and at the same time, it has the advantage of ensuring accessibility and convenience.

However, in the prior art (S900), 1) is operated based on text conversation rather than voice conversation, and 2) an answer including a keyword for a higher-order query must be answered before proceeding to the next step (providing a sub-query menu dependent on the corresponding query) Therefore, it is useful for services in which the process proceeds step-by-step according to a preset procedure, such as a medical appointment service, but is structurally unsuitable for a patient counseling service where the process is not performed step-by-step and that meaningful information must be extracted through dialogue with the patient. It has limitations.

Also, in the prior art S900, only one-way conversation is possible because the next step is possible only when a desired answer is obtained after providing a query by using upper query words and lower query words connected in a tree structure. It has a disadvantage in that it is possible to provide a two-way service to the user, and thus, it has a disadvantage that it is impossible to apply to a medical consultation with a patient.

The present invention is to solve this problem, and the solution of the present invention is that the consultation service management server provides a voice consultation service with the patient, so that the patient needs information such as his or her medical history and symptoms due to the limited medical consultation time in the prior art. The purpose of this study is to provide a doctor-patient consultation support system and method using machine learning natural language processing that can dramatically increase patient satisfaction with treatment by remarkably solving problems that cannot adequately explain the problem.

In addition, another solution to the present invention is that the consultation service management server extracts valid information from the voice consultation with the patient and provides a medical consultation report to the doctor based on the extracted valid information. Machine learning that not only maximizes the effectiveness and efficiency of treatment, but also increases the accuracy and reliability of diagnosis and prescription because it allows more intensive treatment time for patients who need to make medical decisions based on personal information An object of the present invention is to provide a system and method for medical consultation support between doctors and patients using natural language processing.

In addition, another problem to be solved by the present invention is that the doctor-patient conversation, communication, and shared psychology about the disease increase as doctors can use the medical consultation report to have an in-depth conversation with the patient during medical consultation. The purpose of this study is to provide a doctor-patient consultation support system and method using machine learning natural language processing that can dramatically increase the patient's disease treatment rate by increasing the medication compliance of patients.

In addition, another solution to the present invention is that as the consultation service management server provides medical consultation services with patients, the recurrence or exacerbation of the patient is reduced, so that not only can the patient's medical expenses be significantly reduced, but also the patient is limited in time and space The purpose of this study is to provide a system and method for medical consultation support between doctors and patients using machine learning natural language processing that can increase the convenience of treatment by enabling consultation without treatment.

In addition, another solution to the present invention is that the counseling grade setting unit calculates the consumption time (t2) consisting of meaningless information among the patient's voice data, and then selects the consultation grade (L) of the patient according to the calculated consumption time (t2). It is possible to further increase the efficiency and effectiveness of medical consultation by detecting the AI chatbot and proceeding with the difficulty or speed of the conversation differently for each consultation level (L).

The solution of the present invention for solving the above problems is a doctor-patient medical consultation support system comprising a patient-side terminal and a consultation service management server: The service management server includes essential check items for each disease type required for medical consultation and , keywords that can be determined as words related to each essential check item are preset and stored, and by using the patient's corpus and keywords corresponding to the patient's disease type as input data, meaningful information necessary for medical consultation from the patient's corpus a database unit in which a valid information extraction algorithm for extracting valid information is stored; It is executed every preset period (T), and by using the corpus and keywords, learning data that can learn the correlation between the corpus and keywords is generated, and the correlation between the corpus and keywords is performed by using the generated learning data. an artificial intelligence learning unit for learning the valid information extraction algorithm by deriving an extraction model that is a set of parameter values; a voice counseling service unit for performing voice counseling by connecting to the patient-side terminal in response to a request from the patient-side terminal; After recognizing the patient's voice data input from the voice counseling service unit, a corpus is generated, and the patient corpus generated using the effective information extraction algorithm and keywords according to the patient's disease type are input data. An artificial intelligence-based valid information extraction unit for extracting valid information from a patient corpus; a medical consultation report generation unit for generating a medical consultation report based on the valid information extracted by the artificial intelligence-based effective information extraction unit; and a control unit for transmitting the medical consultation report generated by the medical consultation report generating unit to the medical team terminal of the patient and the patient-side terminal, wherein the consultation service management server is executed when the voice consultation is completed by the voice consultation service unit When receiving the patient's voice data executed through the voice counseling service unit, the inputted patient's voice data is analyzed according to a preset grading procedure to set a counseling grade (L) according to the patient's communication ability level , further comprising a counseling grade setting unit stored in the database unit, wherein the voice counseling service unit talks by counseling grade (L) of the patient when the counseling grade (L) of the patient who requested counseling is stored in the database part a data input module for receiving the patient's voice data executed through the voice counseling service unit and the valid information extracted by the artificial intelligence-based valid information extraction unit; The voice data input through the data input module is arranged in a bit string according to the timeline, and the valid information is output section and valid information is output in the voice data arranged in the bit string by referencing and utilizing the input valid information. Voice data classification module for classifying into non-interactive sections; a total output time (t) calculation module for calculating a total output time (t), which is a total output time of the voice data input by the data input module; a consumption time (t2) calculation module for calculating a consumption time (t2) that is an output time that is an output time of the sections in which valid information is not output by the voice data classification module; The consumption rate (R) for calculating the consumption rate (R) which is a percentage of the consumption time (t2) calculated by the consumption time (t2) calculation module and the total output time (t) calculated by the total output time (t) calculation module calculation module; By using the reference table in which the consultation grade (L) is matched for each section of the consumption rate (R), the consultation grade (L) corresponding to the consumption rate (R) calculated by the consumption rate (R) calculation module is detected, and the detected After determining the consultation grade (L) as the consultation grade (L) of the patient, it includes a consultation grade determination module that is stored in the database unit, and the database unit includes a reminder output cycle (T1) for each consultation grade (L). Assuming that the matched second reference table is further stored, and the ventilation cycle output period (T1) of the second reference table is a higher level (or lower level) as the consultation level (L) is higher, the consultation level (L) is set to be proportional to (or inversely proportional to), and the counseling service management server is executed when the counseling grade is determined by the counseling grade setting unit, and by using the second reference table, the After detecting the output cycle (T1) matching the counseling grade (L), the output cycle (T1) detection unit for storing the inspiration output in the database unit; It is executed when the voice consultation by the voice counseling service unit starts, and further comprises a determination unit whether to output a reminder that is terminated when the voice consultation is finished, and the reminder output determination unit searches the database unit at the time of initial execution, A data search and extraction module to search for the ventilation output cycle (T1) for the patient; a monitoring module that is executed when a reminder output period (T1) for the patient is extracted by the data search and extraction module, and monitors the voice counseling service unit and the artificial intelligence-based valid information extraction unit; an elapsed time (Δt) calculation module for calculating, by the monitoring module, an elapsed time (Δt) from the point in time of the question of the query to the current state in which valid information is not extracted by the voice counseling service unit; a comparison module for comparing the elapsed time (Δt) calculated by the elapsed time (Δt) calculation module and the reminder output period (T1) extracted by the data search and extraction module; In the comparison module, if the elapsed time (Δt) is greater than or equal to the reminder output period (T1) (Δt≥T1), the AI chatbot determines that the reminder should be output if the reminder is reached (Δt<T1) , including a reminder output determination module for outputting an output control signal to the voice counseling service unit, and the voice counseling service unit, when an output control signal is input from the reminder output determination unit, disconnects the patient's conversation and repeats the previous query This is to output a reminder that means a repeated comment to ask.

In addition, in the present invention, it is preferable that the voice counseling service unit performs a voice counseling service with a user (patient) of the connected patient-side terminal by utilizing a natural language processing technology (NLP) based on deep learning. do.

In addition, in the present invention, the artificial intelligence-based valid information extracting unit includes: a voice recognition module for recognizing the patient's voice data input from the voice counseling service unit; a corpus generating module for generating a corpus for the patient's voice data recognized by the voice recognition module; a text pre-processing module that pre-processes the atypical text data of the corpus generated by the corpus generating module to suit a purpose before converting it into a feature; a feature vectorization module that converts the text features preprocessed by the text preprocessing module into a vector value that is a numeric value; It is preferable to include an analysis module for extracting valid information from the patient corpus by applying the valid information extraction algorithm to the feature vectorized set by the feature vectorization module.

In addition, in the present invention, a counseling application is installed in the patient-side terminal, and when the counseling application receives a request for counseling from a user (patient), it transmits the counseling request data to the counseling service management server and at the same time transmits the counseling request data to the user (patient) and the counseling It is preferable that a consultation interface, which is a graphic user interface (GUI) providing an interface for the voice consultation service of the service management server, is displayed on the monitor of the patient-side terminal.

In addition, in the present invention, when the counseling application is first executed, when basic information is received from the user (patient), including personal information, disease type, medical staff information who has previously performed medical treatment, and medical staff information who wishes to receive treatment, the input basic information transmits to the consultation service management server, and when the consultation service management server receives basic information from the consultation application, it stores the received basic information in the database unit and at the same time stores the received basic information in the database unit and at the same time, the medical staff who performed the previous treatment included in the basic information or It is preferable to further include a medical team matching unit for matching the medical staff who wish to treat the patient to the medical staff to treat the patient.

In addition, in the present invention, query words for inducing answers related to each essential check item are preset and stored in the database unit, and the consultation application requests a consultation with the consultation service management server according to a consultation request of a user (patient), and , when the voice counseling service unit receives a request for counseling from the counseling application, a data extraction module that searches the database unit to extract the disease type of the patient who requested the consultation, and query words and keywords corresponding to the disease type; a connection module for confirming a connection with a consultation application requesting consultation; It is preferable to further include a voice conversation module for performing voice conversation with a user (patient) of a counseling application connected based on artificial intelligence by using the query words and keywords extracted by the data extraction module.

In addition, in the present invention, the essential check items stored in the database unit are classified into 1) life-related items, which are meaningful items related to life, and 2) disease-related items, which are meaningful items related to diseases. The mandatory check items include at least one or more of meals, sleep, exercise, hobbies, residence, friends, drinking and smoking, and the sub-essential check items constituting the disease-related items are medical history, symptom progress, symptom change, and disease management status. , It is preferable to include at least one or more of a medication state, a psychological state.

delete

delete

delete

According to the present invention having the above problems and solutions, the consultation service management server provides a voice consultation service with the patient, so that the patient cannot sufficiently explain the necessary information such as his/her medical history and symptoms due to the limited medical consultation time in the prior art. By dramatically solving problems, it is possible to dramatically increase patient satisfaction with treatment.

In addition, according to the present invention, after the consultation service management server extracts valid information from the voice consultation with the patient, a medical consultation report is provided to the doctor based on the extracted valid information. Because treatment time can be more intensively devoted to patients who need to make a medical judgment based on this, treatment effectiveness and efficiency can be maximized, and the accuracy and reliability of diagnosis and prescription can be increased.

In addition, according to the present invention, by using the medical consultation report by the doctor, in-depth conversation with the patient is possible during medical consultation, the doctor-patient conversation, communication, and shared psychology about the disease increase, thereby improving the patient's medication compliance. This will dramatically increase the patient's disease treatment rate.

In addition, according to the present invention, as the consultation service management server provides medical consultation service with the patient, the recurrence or exacerbation of the patient is reduced, so that the patient's medical cost can be significantly reduced, and the patient can receive medical consultation without time and space restrictions. This can increase the convenience of treatment.

In addition, according to the present invention, after the consultation grade setting unit calculates the consumption time (t2) consisting of meaningless information among the patient's voice data, according to the calculated consumption time (t2), the consultation grade (L) of the patient is detected, and AI As the chatbot progresses the difficulty or speed of conversation differently for each consultation level (L), the efficiency and effectiveness of medical consultation can be further increased.

1 is a diagram schematically illustrating a method of providing a treatment reservation consultation chatbot service disclosed in Korean Patent Application Laid-Open No. 10-2020-0141957 (Title of the Invention: Chatbot-based method for providing treatment reservation service).
2 is a block diagram illustrating a system for supporting medical consultation between a doctor and a patient according to an embodiment of the present invention.
FIG. 3 is a relationship diagram showing FIG. 2 .
4 is an exemplary diagram illustrating a medical consultation report generated by the consultation service management server of FIG. 2 .
5 is a block diagram illustrating the consultation service management server of FIG. 2 .
FIG. 6 is a conceptual diagram for explaining FIG. 5 .
7 is a block diagram illustrating an AI chatbot-based voice counseling service unit of FIG. 5 .
8 is a block diagram illustrating an artificial intelligence-based valid information extraction unit of FIG. 5 .
9 is a block diagram illustrating a consultation level setting unit of FIG. 5 .
10 is an exemplary view for explaining the total output time (t), the effective time (t1), and the consumption time (t2) of FIG.
11 is a flowchart for explaining an operation process of a system for supporting a doctor-patient consultation according to an embodiment of the present invention.
12 is a flowchart illustrating an AI-based valid information extraction step (S60) of FIG. 11 .
13 is a block diagram illustrating a second embodiment of the consultation service management server of FIG. 2 .
FIG. 14 is a block diagram showing the determination unit whether to output a reminder of FIG. 13 .
Fig. 15 is a flowchart showing the second embodiment of Fig. 11;

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

FIG. 2 is a block diagram illustrating a system for supporting medical consultation between a doctor and a patient according to an embodiment of the present invention, and FIG. 3 is a relational diagram showing FIG. 2 .

The doctor-patient medical consultation support system 1 according to an embodiment of the present invention of FIG. 2 provides a voice counseling service to a patient or guardian, and at the same time provides medically meaningful information such as the patient's medical history, symptom progress, medication status, etc. from voice counseling After extracting the information, a medical consultation report is generated based on the extracted information and provided to the medical staff and patients to maximize the efficiency and effect of medical consultation compared to the same treatment time, as well as the accuracy and reliability of the doctor's diagnosis and prescription. This is to improve patient satisfaction with consultation and medication compliance, and to reduce social medical costs for the treatment of chronically ill patients.

In addition, as shown in FIG. 2, the doctor-patient medical consultation support system 1 of the present invention provides a consultation service to a connected user and at the same time provides meaningful information from voice data through voice consultation with the connected user. After extracting (hereinafter referred to as valid information), the consultation service management server 3 that generates a medical consultation report based on the extracted valid information, the patient-side terminal 5-1, which is a terminal possessed by the patient or patient guardian, ..., (5-N), the medical staff terminal 9 which is a terminal possessed by the doctor, and the patient-side terminal 5-1, ..., are installed in each of (5-N) to manage the counseling service Between the counseling application 7 supporting interworking with the server 3, the counseling service management server 3, the medical staff terminal 9 and the patient side terminal 5-1, ..., (5-N) It consists of a communication network 10 that provides a data movement path.

At this time, in the present invention, for convenience of explanation, the patient-side terminal 5 and the consultation service management server 3 have been described as an example of performing voice consultation through a separate consultation-only application 7, but the patient-side terminal 5 And the voice consultation method of the consultation service management server 3 is not limited thereto, and it is natural that it may be configured to be operated based on a voice call through a normal telephone network.

Also, for convenience of explanation, in the present invention, the AI chatbot uses a machine learning algorithm to perform a consultation service with a patient as an example. can be done through

The communication network 10 supports data communication between the consultation service management server 3, the medical staff terminal 9, and the patient terminal 5-1, ..., (5-N), specifically, the LAN It may be implemented as a (local area network) network, a VAN (Value Added Network) network, a wide area network (WAN), a mobile communication network, Wi-Fi, 3G/4G, LTE, and the like.

The medical staff terminal 9 is a terminal possessed by the medical staff or disposed in the medical staff's office space, and in detail, a desktop-PC (Desktop-PC), a smartphone (Smart phone), a notebook (Notebook), a tablet-PC ( Tablet-PC) and the like. In other cases where it is unavoidable, it is natural that the printed document may also be configured.

In addition, when receiving a medical consultation report for a specific patient from the consultation service management server 3 , the medical staff terminal 9 displays the received medical consultation report according to a request of a user (medical staff).

That is, the medical staff refer to the contents of the medical consultation report transmitted from the consultation service management server 3 through the medical staff terminal 9, and can concentrate on asking questions to hear the information (effective information) required during the limited consultation treatment time. Therefore, it is possible to not only in-depth analysis of the patient but also to maximize the efficiency and effect of counseling by significantly reducing the factors that interfere with the conversation (meaningless conversation) in the contents of the consultation with the patient, thereby increasing the accuracy and effectiveness of diagnosis and prescription. do

The patient-side terminals (5-1), ..., (5-N) are digital devices that a patient or a patient caregiver possesses, and supports connection with the communication network 10, in detail, a smart phone. , a desktop-PC, a notebook, and a tablet-PC.

In addition, the patient-side terminals (5-1), ..., (5-N) are input means 51 such as a keyboard and a touch panel for receiving text from the user, and voice output means such as a speaker from which sound is output ( 52), a voice collecting means 53 such as a microphone that collects voice signals, a display means 54 for displaying characters, numbers, photos, video contents, etc., and communication supporting connection with the communication network 10 It supports voice consultation with the AI chatbot (Chat-bot) of the consultation service management server 3 of FIG. 5 to be described later, including the means 55 .

In addition, a consultation application 7 to be described later is installed in the patient-side terminals 5-1, ..., (5-N).

The consultation application 7 is installed in the patient-side terminal 5 and interlocked with the consultation service management server 3 to provide a voice consultation service with the AI chatbot to the user (patient). Software, application, or It is an application program.

In addition, the counseling application 7 receives basic information from a patient (User) when first executed. In this case, the basic information may include at least one or more of personal information of a patient, a disease type, information on a medical facility and medical staff that performed a previous treatment, and information on a medical facility and medical staff who wish to receive treatment.

In addition, when the counseling application 7 receives basic information from the patient, it transmits the input basic information to the counseling service management server 3 . At this time, the consultation service management server 3 matches the medical staff terminal 9 to perform the treatment of the patient based on the basic information received from the consultation application 7 .

In addition, when the consultation application 7 receives a request for voice consultation from the patient, it transmits the consultation request data to the consultation service management server 3 and at the same time, a graphic for performing voice consultation with the AI chatbot of the consultation service management server 3 A consultation interface, which is a user interface (GUI, Graphic User Interface), is displayed on the display unit 54 of the patient-side terminal 5 . In this case, the consultation interface provides an interface for the voice consultation service between the user (patient or patient guardian) and the AI chatbot of the consultation service management server 3 .

The consultation service management server 3 is a server for providing a consultation service to a patient or patient guardian and, at the same time, providing a medical consultation report containing valid information necessary for medical consultation to a medical staff.

In addition, when the consultation service management server 3 receives basic information from the consultation application 7, it stores the received basic information and at the same time, the contents included in the received basic information (type of disease, medical facilities and medical staff who performed previous treatment) information, the desired medical facility and medical staff information) to match the doctor who treated the patient.

In this case, the matching method may be operated in such a way that the medical facility and medical staff desired by the patient are selected, or the medical facility and the medical staff that the patient has previously treated are selected, and the matching information is stored in the database unit ( 31) is stored.

In addition, when the consultation service management server 3 receives a consultation request from the consultation application 7, it executes an AI-based AI chatbot, and performs voice consultation with the patient through the AI chatbot.

At this time, in the present invention, for convenience of explanation, it has been described as an example that voice counseling with a patient is performed by an AI chatbot. .

The following Table 1 is an exemplary table showing essential check items and query words applied to the AI chatbot of the consultation management server of FIG. 2 .

[Table 1]

The AI chatbot conducts daily conversations with the patient, but as shown in Table 1, the query words for inducing information on the items necessary for the doctor's treatment (hereinafter referred to as essential check items) from the patient are naturally answered according to the situation. ask the

At this time, the essential check items are 1) life-related items that are meaningful items related to life such as meals, sleep, exercise, hobbies, residence, friends, drinking and smoking, and 2) medical history, symptom progress, disease management status, medication compliance, etc. Similarly, it can be classified into higher-order components of disease-related items, which are significant items related to diseases.

As the contents of these essential check items are configured differently according to the patient's symptoms and medical history, the AI chatbot applied to the counseling service management server 3 of the present invention includes essential check items according to the patient's symptoms and medical history and accordingly Query words are preset (configured) and saved.

In this case, the mandatory check items may be set by receiving information from experts and medical staff in advance.

Also, keywords that can be determined as words related to each essential check item are preset and stored in the consultation service management server 3 from the patient's conversation.

At this time, since the mandatory check items may vary depending on the patient's disease type, medical history, and symptoms, the keywords may also have different configurations depending on the patient's disease type, medical history, and symptoms. It may be set by receiving information in advance.

In addition, the counseling service management server 3 learns the effective information extraction algorithm at each preset period (T). At this time, the valid information extraction algorithm is a machine learning algorithm for extracting valid information (keywords and sentences) related to essential check items from the patient's corpus. do it with

In addition, the consultation service management server 3 extracts valid information from the patient's conversation by analyzing the patient's voice data using the previously learned effective information extraction algorithm during voice consultation of the AI chatbot.

In addition, the consultation service management server 3 generates a medical consultation report by using the extracted valid information when the voice consultation is finished, and provides the generated medical consultation report to the medical staff terminal 9 and the patient side terminal 5. . In this case, as the transmission method of the medical consultation report of the consultation service management server 3, various known methods such as encrypted e-mail, text message, dedicated program, and mail transmission may be applied.

4 is an exemplary diagram illustrating a medical consultation report generated by the consultation service management server of FIG. 2 .

As shown in FIG. 4, the medical consultation report 900 lists basic information (personal information, disease type, etc.) of the patient, essential check items, and valid information (patient's answer).

That is, before the actual treatment with the patient, the doctor reads and refers to the medical consultation report 900 transmitted from the consultation service management server 3 to perform actual medical consultation, so the medical consultation report 900 is omitted. By maximizing the effectiveness and efficiency of medical consultation compared to the limited consultation time, it is possible not only to increase the effectiveness of diagnosis and prescription, but also to improve patient satisfaction with treatment. be able to improve

In general, there are cases of patients who answer the doctor's questions during medical consultation, but in the case of patients with related diseases accompanied by cognitive decline such as the elderly or psychiatric medicine, basic understanding and expressiveness are relatively low. For this reason, there is a problem in that a meaningless conversation time (hereinafter referred to as a consumption time t2) increases in the medical consultation time for reasons such as free conversation with a doctor's question or conversation at a very slow speed.

Table 2 below is a table that exemplarily shows a conversation pattern according to the patient's comprehension and expressive ability during medical consultation with a patient with cognitive decline.

[Table 2]

Referring to Table 2, looking at the conversation patterns of patients with cognitive decline, the consumption time (t2) of patients with cognitive decline increases and , in particular, in the case of a patient with low comprehension or expressive power, it can be seen that the consumption time t2 is significantly increased.

That is, the counseling service management server 3 of the present invention provides a voice counseling service with a patient using an AI-based AI chatbot and, at the same time, analyzes the patient's voice data using an effective information extraction algorithm, After extracting valid information (keywords and sentences), a medical consultation report is generated based on the extracted valid information and provided to the medical staff terminal 9 and the patient side terminal 5 .

5 is a block diagram illustrating the consultation service management server of FIG. 2 , and FIG. 6 is a conceptual diagram illustrating FIG. 5 .

As shown in FIGS. 5 and 6 , the counseling service management server 3 includes a control unit 30 , a database unit 31 , a communication interface unit 32 , an artificial intelligence learning unit 33 , and a medical staff matching unit 34 . ), AI chatbot-based voice counseling service unit 35, artificial intelligence-based effective information extraction unit 36, medical consultation report generation unit 37, consultation level setting unit 38, and additional service unit 40.

The control unit 30 is an OS (Operating System) of the consultation service management server 3, and is a control target (31), (32), (33), (34), (35), (36), (37), (38) and (39) are managed and controlled.

In addition, the control unit 30 executes the artificial intelligence learning unit 33 every preset period (T).

In addition, when the control unit 30 receives basic information from the consultation application 7 connected through the communication interface unit 32 , the received basic information is stored in the database unit 31 and inputted to the medical team matching unit 34 at the same time. and the matching information generated by the medical team matching unit 34 is stored in the database unit 31 .

In addition, when the consultation request data is transmitted from the consultation application 7 through the communication interface unit 32 , the control unit 30 executes the AI chatbot-based voice consultation service unit 35 .

In addition, when the voice consultation is terminated by the AI chatbot-based voice consultation service unit 35, the control unit 30 executes the artificial intelligence-based valid information extraction unit 36, and the valid information extracted by the valid information extraction unit 36 The information is input into the medical consultation report generating unit 37 .

In addition, when the medical consultation report is generated by the medical consultation report generating unit 36, the control unit 30 is configured to transmit the generated medical consultation report to the corresponding medical team terminal 9 and the corresponding patient side terminal 5, so that the communication interface unit 32 ) to control

The database unit 31 stores basic information of each patient received from each consultation-only application 7 .

In addition, the database unit 31 stores the matching information of the patient and the medical staff generated by the medical team matching unit 34 .

In addition, as shown in Table 3 below, essential check items A1 for each disease type, query words A2 for each essential check item, and keywords A3 for each disease type are preset and stored in the database unit 31 as shown in Table 3 below. do. In this case, the keywords A3 are defined as words that can be determined to be conversations related to each essential check item.

The following Table 3 is a table exemplarily showing essential check items, query words, and keywords stored in the database unit of the counseling service management server of the present invention.

[Table 3]

In addition, the database unit 31 stores the valid information extraction algorithm learned by the artificial intelligence learning unit 33 .

At this time, the effective information extraction algorithm is learned by the artificial intelligence learning unit 33 every preset period (T), and is utilized in the artificial intelligence-based effective information extraction unit 36 to correspond to the patient's corpus and the patient's disease type. It is a deep-learning algorithm that outputs valid information (keywords and sentences) using keywords as input data.

In addition, the database unit 31 stores an AI chatbot that connects to the consultation application 7 requesting consultation and performs an AI-based voice consultation service with the patient.

This AI chatbot uses a deep-learning algorithm to conduct voice conversations with the connected patient based on artificial intelligence, and at the same time, asks questions about necessary check items and queries according to the patient's disease type to hear valid information from the patient. , and in detail, by using deep-learning algorithms for speech recognition and natural language processing (NLP), the input sentence and output sentence are placed as a pair and the most appropriate expression result is extracted. At this time, Natural Language Processing (NLP) is the five algorithms of word embedding: Continuous Bag-of-Words (CBoW), Relation Network (RN), Convolutional Neural Network (CNN), Self-Attention, It was conducted through a combination of RNN (Recurrent Neural Network), and for Opinion Mining, SentiWordNet Vocabulary Dictionary (assigned numerically as the polarity of words divided into positive, negative, and neutral) and NLTK (Natural Language Toolkit, natural language) A semi-supervised learning algorithm using processing library) was also used.

In addition, the database unit 31 stores the voice data that is the content of the chatbot and the patient's voice consultation performed by the AI chatbot-based voice consultation service unit 35 .

Also, the database unit 31 stores the medical consultation report generated by the medical consultation report generating unit 36 .

In addition, the database unit 31 stores a reference table in which the counseling grade (L) is matched for each consumption rate (R) section. In this case, the consumption rate (R) means the percentage of questions and free conversations occupied by the patient's voice counseling, and will be described in detail with reference to FIGS. 9 and 10 to be described later.

In addition, the database unit 31 stores the consultation level information of the patient set in the consultation level setting unit 38 .

The communication interface unit 32 transmits/receives data to and from the medical staff terminal 9 and the patient-side terminals 5-1, ..., (5-N).

The artificial intelligence learning unit 33 is executed at a preset period (T) under the control of the controller 30 and learns the valid information extraction algorithm utilized in the artificial intelligence-based valid information extraction unit 36 to be described later.

At this time, the valid information extraction algorithm is a machine learning algorithm for extracting valid information, which is a dialogue related to essential check items required for a doctor's treatment, by using the patient's corpus and keywords corresponding to the patient's disease type as input data.

In addition, the valid information extraction algorithm of the artificial intelligence learning unit 33 uses the input patient-side corpus and keywords to generate learning data that can learn the correlation between the patient-side corpus and keywords, and utilizes the generated learning data to generate the patient-side corpus and keywords. An extraction model is derived which is a set of parameter values for correlation between corpus and keywords.

The medical team matching unit 34 receives basic information transmitted from the consultation application 7 through the communication interface unit 32 .

In addition, the medical staff matching unit 34 utilizes the type of disease included in the input basic information, information about medical facilities and medical staff that performed previous treatment, or information on medical facilities and medical staff who wish to receive treatment, and a doctor who will perform treatment for the patient. match the

At this time, the medical staff matching unit 34 matches the medical facilities and medical staff desired by the patient included in the input basic information with the corresponding patient, or matches the medical facilities and medical staff previously treated by the patient included in the input basic information to the corresponding patient. It can be operated in a way that matches with

Also, the matching information generated by the medical team matching unit 34 is stored in the database unit 31 under the control of the control unit 30 .

7 is a block diagram illustrating an AI chatbot-based voice counseling service unit of FIG. 5 .

As shown in FIG. 7 , the AI chatbot-based voice consultation service unit 35 includes a consultation request data input module 351 , a data extraction module 352 , a connection module 353 , and an AI chatbot-based voice chat module 354 . ) is made of

At this time, as described above, the voice counseling service unit 35 of the counseling service management server 3 may be made through a call with a professional counselor.

The consultation request data input module 351 receives consultation request data transmitted from the consultation application 7 through the communication interface unit 32 .

The data extraction module 352 searches the database unit 31 and extracts, from the database unit 31 , the disease type of the patient who requested the consultation, and preset query words and keyword information corresponding to the disease type.

In addition, the data extraction module 352 extracts the consultation grade (L) information when the consultation grade (L) information for the patient is stored in the database unit 31 . At this time, if the patient who requested counseling is a new patient who has not previously performed voice counseling, since a separate counseling grade (L) has not yet been set, the data extraction module 352 separates counseling grade (L) information no need to extract

The connection module 353 confirms the connection with the consultation application 7 in which the AI chatbot requests consultation to the consultation service management server 3 .

The AI chatbot-based voice conversation module 354 performs AI-based voice consultation with the patient to which the AI chatbot is connected.

At this time, as described above, the AI chatbot conducts a voice conversation with the connected patient based on artificial intelligence, but asks the necessary check items and query words according to the patient's disease type to induce a conversation about valid information from the patient. do.

In addition, the AI chatbot can proceed with different difficulty or speed of conversation for each patient's consultation level (L). For example, when the consultation level is low (low), the AI chatbot can slow down the query cycle or easily explain the query.

8 is a block diagram illustrating an artificial intelligence-based valid information extraction unit of FIG. 5 .

The artificial intelligence-based valid information extraction unit 36 of FIG. 8 is executed when the voice consultation is started by the AI chatbot-based voice consultation service unit 35 of FIG. 7 described above, and is terminated when the voice consultation ends.

In addition, the artificial intelligence-based valid information extraction unit 36 extracts valid information from the patient's voice data based on the valid information extraction algorithm learned by the artificial intelligence learning unit 33 described above. In this case, the valid information means meaningful information necessary for doctor's treatment during a patient's conversation, and in detail, it means a keyword for each preset essential check item and sentences highly related to the keyword during the patient's conversation.

In addition, as shown in FIG. 8, the artificial intelligence-based valid information extraction unit 36 includes a voice data input module 361, a voice recognition module 362, a corpus creation module 363, and an AI-based valid information extraction module ( 364).

At this time, the AI-based valid information extraction module 364 includes a text preprocessing module 3641 , a feature vectorization module 3642 , and an analysis module 3643 .

The voice data input module 361 receives voice data from the AI chatbot-based voice conversation module 354 of FIG. 7 described above.

The voice recognition module 362 recognizes the voice by analyzing the input voice data.

The corpus generating module 363 generates a corpus for the patient's voice data recognized by the voice recognition module 362 .

The AI-based effective information extraction module 364 uses the corpus generated by the corpus creation module 363 and the patient's keywords as input data of the valid information extraction algorithm learned by the artificial intelligence learning unit 33, the patient It extracts valid information, which is meaningful information necessary for medical treatment, from the conversation of

In addition, the AI-based valid information extraction module 364 includes a text pre-processing module 3641 , a feature vectorization module 3642 , and an analysis module 3643 .

The text pre-processing module 3641 performs pre-processing according to the purpose before converting the unstructured text data into features, in detail, pre-processing such as tokenization, identity and normalization, stem extraction and lemma extraction, stopword removal, data separation, etc. do.

The feature vectorization module 3642 converts the text feature preprocessed by the text preprocessing module 3641 into a vector value that is a numeric value.

In general, since machine learning algorithms operate by receiving numeric features as data, feature vectorization is required. In this case, for feature vectorization, general-purpose algorithms such as Word2Vec, LDA, FastText, etc. can be used. 1) Word2Vec again has CBOW method and Skip-Gram method. As is widely known at this time, CBOW is a method of predicting a word in the overall context, Skip-gram is a method of inferring surrounding words from a word, and 2) LDA (Latent Dirichlet Allocation) is a process that finds a topic in a set of documents. As a representative algorithm of topic modeling, unsupervised learning and probabilistic algorithms that find hidden topics from a corpus and group topics by keyword are used.

The analysis module 3643 applies the valid information extraction algorithm learned by the artificial intelligence learning unit 33 to the feature vectorized set by the feature vectorization module 3642 to extract valid information from the patient corpus.

The medical consultation report generation unit 37 receives the valid information extracted by the artificial intelligence-based valid information extraction unit 36 .

In addition, the medical consultation report generating unit 37 converts the input valid information according to the format and type of the preset medical consultation report, and then creates the medical consultation report by writing it in the corresponding blank.

At this time, when the medical consultation report is generated by the medical consultation report generating unit 37, the controller 30 transmits the medical consultation report generated to the medical staff terminal 9 matched to the patient by using the matching information of the patient. The communication interface unit 32 is controlled as much as possible. On the other hand, various known methods such as encrypted e-mail, text message, dedicated program, and mail transmission may be applied to the method of providing the medical consultation report of the consultation service management server 3 .

FIG. 9 is a block diagram illustrating the consultation level setting unit of FIG. 5 , and FIG. 10 is an exemplary diagram for explaining the total output time t, the effective time t1, and the consumption time t2 of FIG. 9 .

As shown in FIG. 9, the consultation level setting unit 38 calculates the data input module 381, the voice data classification module 382, the total output time (t) calculation module 383, and the effective time (t1). It consists of a module 384 , a consumption time t2 calculation module 385 , a consumption rate (R) calculation module 386 , and a consultation grade determination module 387 .

The data input module 381 receives the voice data of the voice consultation executed by the AI chatbot-based voice consultation service unit 35 and the valid information extracted by the AI-based valid information extraction unit 36 .

The voice data classification module 382 aligns the input voice data into a bit string according to the timeline, and refers to and utilizes the input valid information to validate the voice data arranged in the bit string as shown in FIG. 10 . It is classified into a section in which information is output and a section in which valid information is not output.

The total output time (t) calculation module 383 calculates the total output time (t), which is the total output time of the voice data.

The effective time (t1) calculation module 384 calculates the effective time (t1), which is an output time obtained by summing the sections in which valid information is output by the voice data classification module 382 .

The consumption time (t2) calculation module 385 calculates the consumption time t2, which is an output time obtained by summing the sections in which valid information is not output by the voice data classification module 382 .

The consumption rate (R) calculation module 386 includes the consumption time t2 calculated by the calculation module 385 and the total output time t calculated by the consumption time t2 and the total output time t calculation module 383 . ), which is a percentage of the consumption rate (R) is calculated.

That is, the more the patient engages in unnecessary conversations unrelated to the question during voice counseling, the higher the consumption rate (R).

The consultation grade determination module 387 searches the reference table stored in the database unit 31, and detects the consultation grade (L) corresponding to the consumption rate (R) calculated by the consumption rate (R) calculation module 386, The detected counseling grade (L) is determined as the counseling grade (L) of the patient.

In this case, the reference table is data in which the counseling grade (L) is matched for each section of the consumption rate (R).

Also, the consultation level L determined by the consultation level determination module 387 is stored in the database unit 31 under the control of the controller 30 .

The additional service unit 39 performs other additional services necessary for the operation of the consultation service management server 3 , such as treatment reservation, civil complaint processing, and consultation application 7 update.

11 is a flowchart for explaining an operation process of a system for supporting a doctor-patient consultation according to an embodiment of the present invention.

As shown in FIG. 11, the doctor-patient medical consultation support method (S1) of the present invention includes an artificial intelligence learning step (S10), a medical team matching step (S20), a consultation request step (S30), and an AI chatbot connection step (S40), AI chatbot-based voice consultation step (S50), artificial intelligence-based valid information extraction step (S60), medical consultation report generation step (S70), consultation grade setting step (S80), medical consultation report transmission step (S90) is made of

In the artificial intelligence learning step (S10), the counseling service management server 3 uses the patient corpus and keywords as input data, and learns the valid information extraction algorithm, which is a deep-learning algorithm for outputting valid information, every preset period (T). This is the stage of reconciliation.

At this time, the artificial intelligence learning step (S10) uses the patient-side corpus and keywords collected during the period (T) to generate learning data that can learn the correlation between the patient-side corpus and keywords, and utilize the generated learning data to generate the patient-side corpus An extraction model is derived which is a set of parameter values for correlation between corpus and keywords.

In addition, the effective information extraction algorithm learned by the artificial intelligence learning step (S10) is utilized in the artificial intelligence-based effective information extraction step (S60).

In the medical team matching step (S20), when the consultation service management server 3 receives basic information from the consultation application 7, the received basic information is used to match the doctor who treated the patient.

In addition, in the medical staff matching step (S20), matching information of the matched patient and the medical staff terminal is stored in the database unit 31 .

The consultation request step (S30) is a step in which the consultation application 7 transmits the consultation request data to the consultation service management server 3 according to the request of the user (patient or patient guardian).

The AI chatbot connection step ( S40 ) is a step in which the AI chatbot of the consultation service management server 3 connects with the consultation application 7 .

The AI chatbot-based voice consultation step (S50) is a step in which the AI chatbot conducts an AI-based voice consultation service with the patient through the consultation application (7).

12 is a flowchart illustrating an AI-based valid information extraction step (S60) of FIG. 11 .

In the AI-based valid information extraction step (S60), as shown in FIG. 12, the consultation service management server 3 analyzes the voice data input from the AI chatbot-based voice consultation step (S50) to recognize a voice. Step (S61), the consultation service management server (3) using the recognition information detected by the voice recognition step (S61) to generate a corpus (Corpus) generating a corpus (S62), and the consultation service management server (3) ) of the text pre-processing step (S63) of pre-processing before converting the unstructured text data of the corpus generated by the corpus creation step (S62) into features, and the text pre-processing step (S63). Feature vectorization step (S64) of converting a value into a vector value, and the effective information extraction algorithm learned in the artificial intelligence learning step (S10) in the feature vectorized set by the counseling service management server 3 by the feature vectorization step (S64) By applying , it consists of an analysis and extraction step (S65) of extracting valid information from the patient corpus.

Referring back to FIG. 11 and looking at the medical consultation report generation step (S70), the medical consultation report generation step (S70) is performed by the consultation service management server 3 by the artificial intelligence-based valid information extraction step (S60) of FIG. 11 described above. This is a step of generating a medical consultation report by converting the extracted valid information according to a preset format and type of the medical consultation report, and then writing it in a corresponding blank.

The consultation grade setting step (S80) is performed by using the voice data of the voice consultation conducted in the AI chatbot-based voice consultation step (S50) and the valid information extracted by the AI-based effective information extraction step (S60), After calculating the total output time (t), the effective time (t1), and the consumption time (t2), the consumption rate (R) is calculated using these.

In addition, the counseling grade setting step (S80) utilizes the calculated consumption rate (R) and the reference table to determine the counseling grade (L) of the patient, and stores the determined counseling grade (L) in the database unit (31).

In the medical consultation report transmission step (S90), the consultation service management server 3 transmits the medical consultation report generated by the medical consultation report generation step (S70) to the consultation application 7 of the patient and the medical team terminal matching the patient ( 9) is the transmission step.

The medical consultation support system 1 for a doctor-patient according to an embodiment of the present invention configured as described above provides a voice counseling service to a patient or guardian and provides medically meaningful information from voice counseling to the patient's medical history, symptom progress, medication status, etc. After extracting information, a medical consultation report is generated based on the extracted information and provided to medical staff and patients to maximize the efficiency and effect of medical consultation compared to the same treatment time, as well as to improve the accuracy and accuracy of doctor diagnosis and prescription. Reliability increases, patient satisfaction with consultation and medication compliance are improved, and it has the advantage of reducing social medical costs for the treatment of chronic disease patients.

13 is a block diagram illustrating a second embodiment of the consultation service management server of FIG. 2 .

In the second consultation service management server 23 of FIG. 13, according to the patient's consultation grade (L), the voice consultation of the AI chatbot is optimally operated, thereby improving the efficiency of the voice consultation of the AI chatbot and reducing the consumption time (t2) it is to make

In addition, the second consultation service management server 23 is, as shown in FIG. 13, the communication interface unit 32, the artificial intelligence learning unit 33, the medical staff matching unit 34, and the AI-based validity of FIG. It includes an information extraction unit 36 , a medical consultation report generation unit 37 , a consultation level setting unit 38 , and an additional service unit 39 .

In addition, the second consultation service management server 23 includes a second control unit 230 , a second database unit 231 , a second AI chatbot-based voice consultation service unit 235 , a reminder output determination unit 236 , and ventilation It further includes a comment output period (T1) detection unit (240).

The second control unit 230 performs the same operation as the control unit 30 of FIG. 5 described above, but when the ventilation output cycle (T1) is detected by the ventilation output cycle (T1) detection unit 240, the detected ventilation The comment output period T1 is stored in the second database unit 231 .

The second database unit 231 stores the same data as the database unit 30 of FIG. 5 described above, and a second reference table in which the reminder output period T1 is matched for each consultation grade (L) is further stored.

At this time, the reminder message means a repeated message for the AI chatbot to naturally cut off the patient's conversation and ask the question again, and the reminder message output cycle (T1) means the maximum value of the continuous consumption time (t2).

In addition, assuming that the higher the consultation level (L), the higher the level, the ventilation cycle output period (T1) of the second reference table is set to be proportional to the consultation level (L). That is, the ventilation cycle output cycle (T1) of the second reference table is set longer as the consultation level (L) is higher, and set shorter as the consultation level (L) is lower.

For example, assuming that the higher the counseling grade (L), the higher the level, and the level of the counseling grade (L) is classified from '1' to '10', having an excellent counseling grade (L) of level 10 The reminder output period (T1) of the patient (A) is set to be long, and the reminder output period (T1) of the patient (B) having a low counseling grade (L) of level 1 may be set to be short.

In general, in the case of patients with cognitive decline, such as old age or mental illness, the consumption time (t2), which is meaningless during voice counseling, increases, and this characteristic not only reduces the effectiveness of counseling but also the entire conversation. This causes a problem in that unnecessary computational throughput of processors increases due to an increase in time. The present invention is to solve this problem, and is configured to detect the counseling grade calculated based on the patient's previous voice counseling and the reminder output period (T1) according to the counseling grade (L), and based on the second AI chatbot When the voice counseling service unit 235 is configured to output a reminder when the patient's consumption time t2 continues, it can naturally evoke the patient's concentration and lead to an answer to the desired question, as well as the consumption time t2. can be significantly reduced.

The second AI chatbot-based voice consultation service unit 235 performs the same operation as the AI chatbot-based voice consultation service unit 35 of FIG. 5 described above, but from the reminder output determination unit 236 of FIG. When the output control signal is input, preset voice data of a reminder is output to the patient.

At this time, since the length of the reminder output period (T1) is determined according to the consultation grade (L) according to the patient's previous consultation history, it is possible to significantly reduce the consumption time (t2) of the patient.

FIG. 14 is a block diagram showing the determination unit whether to output a reminder of FIG. 13 .

The output determination unit 236 of FIG. 14 is executed when the voice consultation of the AI chatbot is started through the second AI chatbot-based voice consultation service unit 235 .

In addition, as shown in FIG. 14, the reminder output determination unit 236 includes a data search and extraction module 2361, a monitoring module 2362, an elapsed time (Δt) calculation module 2363, a comparison module ( 2364), it consists of a ventilation ment output determination module (2365).

The data search and extraction module 2361 searches the second database unit 231, and searches for a reminder output period (T1) for the patient.

In addition, the data search and extraction module 2361 1) when the ventilation sentence output cycle (T1) for the patient is searched, extracts the corresponding ventilation sentence output cycle (T1), 2) the ventilation output cycle for the patient ( T1) is not searched, and the patient is judged as a new patient, and a separate operation is not performed thereafter.

The monitoring module 2362 is executed when the reminder output period T1 for the patient is extracted by the data search and extraction module 2361, and the second AI chatbot-based voice consultation service unit 235 and artificial intelligence-based The valid information extraction unit 36 is monitored.

When monitoring by the monitoring module 2362, the elapsed time (Δt) calculation module 2363 calculates the elapsed time (Δt) from the time when the AI chatbot asks a query to the present in which valid information is not extracted continuously. Calculate.

The comparison module 2364 compares the elapsed time (Δt) calculated by the elapsed time (Δt) calculation module 2363 and the reminder output period (T1) extracted by the data search and extraction module 2361 And in detail, it is compared (Δt≥T1) whether the elapsed time (Δt) reaches the ventilation impression output period (T1).

In the comparison module 2364, the reminder output determination module 2365 1) If the elapsed time (Δt) is less than the reminder output period (T1) (Δt < T1), the AI chatbot has not yet output the reminder 2) If the elapsed time (Δt) is greater than or equal to the reminder output period (T1) (Δt≥T1), it is determined that the AI chatbot should output the reminder.

In addition, the reminder output determination module 2362 outputs an output control signal to the second AI chatbot-based voice consultation service unit 235 when determining the output of the reminder message of the AI chatbot.

At this time, when the second AI chatbot-based voice counseling service unit 235 receives an output control signal from the reminder output determination unit 236, it outputs preset voice data.

Returning to FIG. 13 again, looking at the ventilation ment output cycle (T1) detection unit 240, the ventilation ment output cycle (T1) detection unit 240 is the consultation grade of the patient set by the consultation grade setting unit 38 (L) Upon receiving the input, the second reference table stored in the database unit 231 is referenced and utilized, and a reminder output period (T1) matching the consultation grade (L) of the patient is detected.

At this time, the reminder output period (T1) is, as described above, the higher the counseling grade (L), the higher the level, the higher the counseling grade (L), the longer the length is set, and the lower the counseling grade (L) is set shorter.

Fig. 15 is a flowchart showing the second embodiment of Fig. 11;

The second doctor-patient medical consultation support method (S200) of FIG. 15 includes an artificial intelligence learning step (S210), a medical team matching step (S220), a consultation request step (S230), an AI chatbot connection step (S240), the second AI chatbot-based voice consultation step (S250), artificial intelligence-based valid information extraction step (S260), monitoring step (S270), voice consultation end determination step (S280), evoking message output determination step (S290), output control signal It consists of an output step (S300), a medical consultation report generation step (S310), a consultation grade setting step (S320), a reminder output cycle (T1) detection step (S330), and a medical consultation report transmission step (S340).

At this time, the artificial intelligence learning step (S210) of FIG. 15, the medical staff matching step (S220), the consultation request step (S230), the AI chatbot connection step (S240), the AI-based valid information extraction step (S260), and a medical consultation report generation Step (S310), consultation level setting step (S320), medical consultation report transmission step (S340), the artificial intelligence learning step (S10) of FIG. 11, the medical staff matching step (S20), the consultation request step (S30) and , AI chatbot access step (S40), AI-based valid information extraction step (S60), medical consultation report generation step (S70), consultation level setting step (S80), medical consultation report transmission step (S90) Therefore, a detailed description will be omitted.

The second AI chatbot-based voice consultation step (S250) proceeds in the same manner as the AI chatbot-based voice consultation step (S50) of FIG. Upon receiving the output control signal from the output determination step (S290), it outputs preset voice data of a reminder.

The monitoring step (S270) is a step of monitoring the AI chatbot-based voice consultation step (S250) and the artificial intelligence-based valid information extraction step (S260).

In addition, the monitoring step (S270) searches the second database unit 231, and searches for a reminder output period (T1) for the patient.

In addition, the monitoring step (S270) calculates the elapsed time (Δt) from the point in time when the AI chatbot makes a query in the AI chatbot-based voice consultation step (S250) to the present in which valid information is not extracted continuously.

The voice consultation end determination step (S280) is a step of determining whether the voice consultation has ended in the AI chatbot-based voice consultation step (S250) through the monitoring step (S270).

In addition, the voice consultation end determination step (S280) is 1) if the voice consultation of the AI chatbot is finished, the next step is to generate a medical consultation report (S310) and set the consultation level (S320), and 2) if the AI If the chatbot's voice consultation is in progress, the next step proceeds to the determination step (S290) of whether to output a reminder.

The determination step (S290) of whether to output a reminder is carried out when it is determined that the voice consultation of the AI chatbot is in progress in the determination step (S280) of whether to end the voice consultation, and the elapsed time (Δt) calculated in the monitoring step (S270) and the corresponding Compare the ventilation moment output period (T1) of the patient.

In addition, the ventilatory output determination step (S290) is 1) if the elapsed time (Δt) is less than (Δt < T1) of the ventilation impression output period (T1), proceed to the monitoring step (S270) to the next step, 2) If the elapsed time (Δt) is greater than or equal to the ventilation ment output period (T1) (Δt≥T1), the output control signal output step (S300) proceeds to the next step.

The output control signal output step (S300) proceeds when the elapsed time (Δt) is greater than or equal to the ventilation message output period (T1) (Δt ≥ T1) in the ventilation message output determination step (S290), and output control with an AI chatbot This is the step of outputting the signal.

The reminder output cycle (T1) detection step (S330) is performed after the consultation service management server 3 is the consultation grade determination step (S320), and searches the preset second reference table, the consultation grade determination step (S320) It is a step of detecting a reminder output period (T1) matched to the consultation grade (L) of the patient determined by, and storing the detected reminder output period (T1) in the database unit.

In the doctor-patient medical consultation support system according to an embodiment of the present invention configured as described above, the consultation service management server provides a voice consultation service with the patient, so that the patient can save his/her medical history and symptoms due to the limited medical consultation time in the prior art. By remarkably solving the problem of not sufficiently explaining the necessary information, it is possible to dramatically increase the patient's satisfaction with the treatment.

In addition, the doctor-patient consultation support system of the present invention provides a medical consultation report to the doctor based on the extracted valid information after the consultation service management server extracts valid information from the voice consultation with the patient. By referring to the in-depth information, more intensive treatment time can be devoted to patients who need to make medical decisions based on in-depth information, so it is possible not only to maximize the effectiveness and efficiency of treatment, but also to improve the accuracy and reliability of diagnosis and prescription. can be raised

In addition, the doctor-patient consultation support system of the present invention enables a doctor to have an in-depth conversation with a patient during medical consultation by using a medical consultation report. It is possible to dramatically increase the patient's disease treatment rate by increasing the patient's medication compliance.

In addition, in the doctor-patient consultation support system of the present invention, as the consultation service management server provides a consultation service with the patient, the recurrence or exacerbation of the patient is reduced, thereby significantly reducing the patient's treatment cost and saving the patient's time. And medical consultations are possible without space restrictions, so it is possible to increase the convenience of treatment.

In addition, in the doctor-patient consultation support system of the present invention, after the consultation grade setting unit calculates the consumption time (t2) consisting of meaningless information among the patient's voice data, the consultation grade of the corresponding patient according to the calculated consumption time (t2) (L) is detected and the AI chatbot proceeds with different difficulty or speed of conversation for each consultation grade (L), thereby further enhancing the efficiency and effectiveness of medical consultation.

1: Doctor-patient consultation support system
3: Counseling service management server 5-1, ..., 5-N: Patient side terminals
7: Consultation application 9: Medical staff terminal
10: communication network 30: control unit
31: database unit 32: communication interface unit
33: artificial intelligence learning unit 34: medical staff matching unit
35: AI chatbot-based voice counseling service unit 36: AI-based valid information extraction unit
37: medical consultation report generation unit 38: consultation level setting unit
39: additional service unit 51: input means
52: audio output means 53: audio collection means
54: display means 55: communication means
351: consultation request data input module 352: data extraction module
353: access module 354: AI chatbot-based voice conversation module
361: voice data input module 362: voice recognition module
363: corpus generation module 364: AI-based valid information extraction module
381: data input module 382: voice data classification model
383: Total output time (t) calculation module 384: Effective time (t1) calculation module
385: consumption time (t2) calculation module 386: consumption rate (R) calculation module
387: Counseling grade determination module

Claims (10)

In a doctor-patient medical consultation support system comprising a patient-side terminal and a consultation service management server:
The service management server
Essential check items for each disease type required for medical consultation and keywords that can be determined as words related to each essential check item are preset and stored, and the patient's corpus and keywords corresponding to the patient's disease type are used as input data. , a database unit in which a valid information extraction algorithm for extracting valid information, which is meaningful information necessary for medical consultation, is stored from the patient's corpus;
It is executed every preset period (T), and by using the corpus and keywords, learning data that can learn the correlation between the corpus and keywords is generated, and the correlation between the corpus and keywords is performed by using the generated learning data. an artificial intelligence learning unit for learning the valid information extraction algorithm by deriving an extraction model that is a set of parameter values;
a voice counseling service unit for performing voice counseling by connecting to the patient-side terminal in response to a request from the patient-side terminal;
After recognizing the patient's voice data input from the voice counseling service unit, a corpus is generated, and the patient corpus generated using the effective information extraction algorithm and keywords according to the patient's disease type are input data. An artificial intelligence-based valid information extraction unit for extracting valid information from a patient corpus;
a medical consultation report generation unit for generating a medical consultation report based on the valid information extracted by the artificial intelligence-based effective information extraction unit;
a control unit for transmitting the medical consultation report generated by the medical consultation report generating unit to a medical team terminal of the patient and the patient side terminal,
The consultation service management server
It is executed when the voice counseling is completed by the voice counseling service unit, and when the voice data of the patient executed through the voice counseling service unit is received, the inputted patient's voice data is analyzed according to a preset grading procedure to communicate with the patient. After setting the consultation level (L) according to the ability level, further comprising a consultation level setting unit for storing in the database unit,
The voice counseling service department
When the counseling grade (L) of the patient who requested counseling is stored in the database unit, the difficulty or speed of the conversation is different for each counseling grade (L) of the patient,
The consultation level setting unit
a data input module for receiving the patient's voice data executed through the voice counseling service unit and the valid information extracted by the artificial intelligence-based valid information extraction unit;
The voice data input through the data input module is arranged in a bit string according to the timeline, and the valid information is output section and valid information is output in the voice data arranged in the bit string by referencing and utilizing the input valid information. Voice data classification module for classifying into non-interactive sections;
a total output time (t) calculation module for calculating a total output time (t), which is a total output time of the voice data input by the data input module;
a consumption time (t2) calculation module for calculating a consumption time (t2) that is an output time that is an output time of the sections in which valid information is not output by the voice data classification module;
The consumption rate (R) for calculating the consumption rate (R) which is a percentage of the consumption time (t2) calculated by the consumption time (t2) calculation module and the total output time (t) calculated by the total output time (t) calculation module calculation module;
By using the reference table in which the consultation grade (L) is matched for each section of the consumption rate (R), the consultation grade (L) corresponding to the consumption rate (R) calculated by the consumption rate (R) calculation module is detected, and the detected After determining the consultation grade (L) as the consultation grade (L) of the patient, including a consultation grade determination module for storing in the database unit,
A second reference table is further stored in the database unit in which the output period (T1) of the reminder output is matched for each consultation grade (L),
The ventilation cycle output cycle (T1) of the second reference table is set to be proportional (or inversely proportional) to the counseling grade (L), assuming that the higher the counseling grade (L), the higher the level (or the lower level),
The consultation service management server
It is executed when the consultation grade is determined by the consultation grade setting unit, and by using the second reference table, a reminder output period (T1) matching the consultation grade (L) of the patient determined by the consultation grade setting unit After detection, a reminder output period (T1) detection unit to store in the database unit;
It is executed when the voice counseling by the voice counseling service unit is started, further comprising a determination unit whether to output a reminder that is terminated when the voice counseling is finished,
The ventilation comment output determination unit
a data search and extraction module for searching the database unit when first executed, and searching for a reminder output period (T1) for the patient;
a monitoring module that is executed when a reminder output period (T1) for the patient is extracted by the data search and extraction module, and monitors the voice counseling service unit and the artificial intelligence-based valid information extraction unit;
an elapsed time (Δt) calculation module for calculating, by the monitoring module, an elapsed time (Δt) from the point in time of the question of the query to the current state in which valid information is not extracted by the voice counseling service unit;
a comparison module for comparing the elapsed time (Δt) calculated by the elapsed time (Δt) calculation module and the reminder output period (T1) extracted by the data search and extraction module;
In the comparison module, if the elapsed time (Δt) is greater than or equal to the reminder output period (T1) (Δt≥T1), the AI chatbot determines that the reminder should be output if the reminder is reached (Δt<T1) , including a reminder output determination module for outputting an output control signal to the voice counseling service unit,
The voice counseling service department
When an output control signal is input from the ventilating sentence output determination unit, the doctor-patient consultation support system, characterized in that the ventilating sentence is output, which means a repeated sentence for stopping the patient's conversation and asking the previous query again. The method of claim 1, wherein the voice counseling service unit
Medical consultation support for doctor-patients, characterized in that voice consultation services are performed with the user (patient) of the connected patient-side terminal using natural language processing (NLP) based on deep learning system. The method of claim 2, wherein the artificial intelligence-based valid information extraction unit
a voice recognition module for recognizing the patient's voice data input from the voice counseling service unit;
a corpus generating module for generating a corpus for the patient's voice data recognized by the voice recognition module;
a text pre-processing module that pre-processes the unstructured text data of the corpus generated by the corpus generating module into a feature before converting it into a feature;
a feature vectorization module that converts the text features preprocessed by the text preprocessing module into a vector value that is a numeric value;
and an analysis module for extracting valid information from a patient corpus by applying the valid information extraction algorithm to the feature vectorized set by the feature vectorization module. According to claim 3, wherein the patient-side terminal is installed with a counseling application,
The counseling application is
When receiving a consultation request from a user (patient), a graphical user interface (GUI) that transmits the consultation request data to the consultation service management server and provides an interface for the voice consultation service of the user (patient) and the consultation service management server at the same time , Graphic User Interface), a doctor-patient consultation support system, characterized in that the display of the consultation interface on the monitor of the patient-side terminal. The method of claim 4, wherein the counseling application is
Upon initial execution, when basic information is received from the user (patient), including personal information, disease type, information on medical staff who performed previous treatment, and information on medical personnel who wish to receive treatment, the input basic information is transmitted to the counseling service management server do,
The consultation service management server
Upon receiving the basic information from the counseling application, the received basic information is stored in the database unit and at the same time, the medical staff who performed the previous treatment included in the basic information or the medical staff desiring treatment are matched with the medical staff to treat the patient. Doctor-patient consultation support system, characterized in that it further comprises a matching unit. [6] The method of claim 5, wherein query words for inducing answers related to each essential check item are preset and stored in the database unit,
The counseling application requests counseling to the counseling service management server according to the counseling request of the user (patient),
The voice counseling service department
a data extraction module that, upon receiving a request for consultation from the consultation application, searches the database unit, and extracts the disease type of the patient who requested the consultation, and query words and keywords corresponding to the disease type;
a connection module for confirming a connection with a consultation application requesting consultation;
Physician-patient, characterized in that it further comprises a voice conversation module for performing voice conversation with the user (patient) of the consultation application connected based on artificial intelligence by using the query words and keywords extracted by the data extraction module. Medical consultation support system. The method of claim 6, wherein the essential check items stored in the database unit are classified into 1) a life-related item that is a meaningful item related to life, and 2) a disease-related item that is a meaningful item related to a disease,
Sub-essential check items constituting the life-related items include at least one or more of meals, sleep, exercise, hobbies, residence, friends, drinking and smoking,
The sub-essential check items constituting the disease-related items include at least one of medical history, symptom progress, symptom change, disease management status, medication status, and psychological status. Doctor-patient consultation support system. delete delete delete

Images