KR20230110200A - Method of locating for optimizing examination sequence using AI learning model and System using the same method - Google Patents

Abstract

The checkup order assignment method of the present invention includes: a checkup data collection step (S100) of collecting patient characteristic data according to characteristics of the patient and a checkup time required for each checkup item in the data collection unit 100; a checkup time model learning step (S200) of learning the patient characteristic data and checkup time in the model learning unit 200 and generating a model capable of predicting the checkup time according to the patient characteristic data; and a waiting number calculation unit 300 , a step of determining the number of waiting patients for each checkup item (S300), a checkup order allocating step (S400) of allocating optimized checkup items to patients in real time in the checkup order assigning unit 400, and a time information providing step of providing waiting time and final completion time to patients according to patient characteristic data and the number of waiting patients in the time calculation unit 500 (S500).

Description

Method of locating for optimizing examination sequence using AI learning model and System using the same method}

The present invention relates to a method for assigning a checkup order using an artificial intelligence learning model and a system using the same, and more specifically, to a method for assigning a checkup order using an artificial intelligence learning model in which a learning model is applied and automatically optimized, and a system using the same.

In general, the order of examinations in health examination facilities uses a pre-determined order. Instead, there is no way to solve this problem when the local schedule is changed, such as a phenomenon in which the patient's characteristics or each examination item are concentrated.

Since the patients move in the order in which they have already been designated, when each checkup item is finished, the next checkup item moves to a designated position. In the normal case, since they move in the designated order, the individual process proceeds linearly for all patients unless there is a special change.

For example, if an eye test takes a lot of time, a bottleneck occurs in the eyesight test checkup item. For this reason, inefficiency occurs, such as the number of patients undergoing examination is insufficient in the examination items performed after the vision examination, and the overall examination time is increased.

Therefore, it is necessary to develop a specific method for efficiently using these health checkup facilities by considering the characteristics of individual patients and the processing status of each checkup item.

Korean Registered Patent No. 2164202

The present invention has been made to solve the above conventional problems, and an object of the present invention is to provide a checkup order assignment method using an artificial intelligence learning model and a system using the same.

The checkup order assignment method of the present invention includes: a checkup data collection step (S100) of collecting patient characteristic data according to characteristics of the patient and a checkup time required for each checkup item in the data collection unit 100; a checkup time model learning step (S200) of learning the patient characteristic data and checkup time in the model learning unit 200 and generating a model capable of predicting the checkup time according to the patient characteristic data; and a waiting number calculation unit 300 , a step of determining the number of waiting patients for each checkup item (S300), a checkup order allocating step (S400) of allocating optimized checkup items to patients in real time in the checkup order assigning unit 400, and a time information providing step of providing waiting time and final completion time to patients according to patient characteristic data and the number of waiting patients in the time calculation unit 500 (S500).

In one embodiment, the waiting number calculation unit 300 includes a beacon unit 310, and the beacon unit 310 may be characterized in that it identifies the number of people waiting for each item in association with the patient's smartphone.

In one embodiment, the checkup item assignment step (S400) is a timeline construction step (S410) of predicting and constructing a timeline including start time, end time, and waiting time for each checkup item for all of a plurality of patients by using the checkup order assignment unit 400. An assignment exchange step (S420) of exchanging a checkup item and a second checkup item of another random patient (S420), a repeat execution step (S430) of repeating the exchange assignment step (S420) in real time based on the checkup item for a random patient if one checkup item for a random patient ends differently from the predicted checkup time (S430), and a checkup item assignment step (S440) of allocating the next checkup item optimized for the end point of the patient who has completed one checkup item. can be characterized.

In one embodiment, in the time line construction step (S410), if the end time of another checkup item is later than the start time of the patient's individual checkup item, the start time of one checkup item may be shifted to after the end time of the other checkup item.

In one embodiment, the step of constructing a timeline (S410) may further include a process of predicting and constructing a timeline including checkup start times, end times, and waiting times for all patients for all a plurality of checkup items.

The checkup order assignment system of the present invention includes: a data collection unit 100 that collects patient characteristic data according to patient characteristics and checkup time required for each checkup item; a model learning unit 200 that learns patient characteristic data and checkup time and generates a model capable of predicting the checkup time according to patient characteristic data; and a time calculation unit 500 that provides waiting time and final completion time to the patient according to patient characteristic data and the number of waiting patients.

In one embodiment, the waiting number calculation unit 300 includes a beacon unit 310, and the beacon unit 310 may be characterized in that it identifies the number of people waiting for each item in association with the patient's smartphone.

In one embodiment, the checkup order assigning unit 400 is a timeline component that predicts and constructs a timeline including start time, end time, and waiting time for each checkup item for a plurality of patients, and an assignment exchange unit that exchanges the first checkup item of a patient and the second checkup item of another patient when the final completion time of a patient or another patient decreases when the first checkup item of a patient and the second checkup item of another patient are exchanged with each other. , If one checkup item for a certain patient ends differently than the estimated time required for checkup, a repeat implementation unit that repeats the exchange assignment step (S420) in real time centering on the checkup item for the random patient, and a checkup item allocator that allocates the next checkup item optimized for the end point of the patient who has completed one checkup item.

In one embodiment, the time line configuration unit shifts the start time of one checkup item to an end time of another checkup item when the end time of another checkup item is later than the start time of the patient's individual checkup item.

In one embodiment, the timeline configuration unit may be characterized in predicting and constructing a timeline including a checkup start time, end time, and waiting time for all patients for all a plurality of checkup items.

Therefore, according to the present invention, it is possible to predict the examination time required for a specific examination item according to the characteristics of the patient.

Based on the predicted checkup time, a time table for each patient and each checkup item can be precisely simulated, so that the optimal checkup time can be assigned accordingly.

In addition, even when the simulated checkup time is changed, by continuously and repeatedly changing the checkup order for each patient under better conditions in real time, such optimization continues until just before the patient actually enters the checkup, so that the patient and the hospital can perform the checkup while spending the optimal time.

1 is a flowchart illustrating a checkup order assignment method using an artificial intelligence learning model according to an embodiment of the present invention.
FIG. 2 is a schematic configuration diagram illustrating a checkup order assignment system using an artificial intelligence learning model according to an embodiment of FIG. 1 .
3 is a flowchart illustrating a step of allocating checkup items in a checkup order assignment method using an artificial intelligence learning model according to another embodiment of the present invention.
4A to 4B are data sets for explaining a checkup order allocation method using the artificial intelligence learning model of FIG. 1 .
5A to 5B are timeline sets for explaining a checkup order assignment method using the artificial intelligence learning model of FIG. 1 .

Hereinafter, embodiments disclosed herein will be described with reference to the accompanying drawings.

In particular, even if there are no reference numerals, components that can be inferred from the description in the specification can be interpreted according to the literal description.

Even words used at the level of those skilled in the art may be misused because they do not accurately reflect actual functions and roles. In this case, it is a principle to infer based on the contents uniformly described in the specification and to be interpreted practically according to the intention of the inventor, and it is not necessary to limit the interpretation to the limitation of the text itself.

The embodiments disclosed in this specification will be described in detail with reference to the accompanying drawings, but the same or similar elements are assigned the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted.

The suffixes "module" and "unit" for components used in the following description are given or used together in consideration of ease of writing the specification, and do not have meanings or roles that are distinct from each other by themselves. In addition, in describing the embodiments disclosed in this specification, if it is determined that a detailed description of a related known technology may obscure the gist of the embodiment disclosed in this specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in this specification, and the technical idea disclosed in this specification is not limited by the accompanying drawings, and all changes included in the spirit and technical scope of the present invention It should be understood to include equivalents or substitutes.

It should be understood that when an element is referred to as being “connected” or “connected” to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle.

Singular expressions include plural expressions unless the context clearly dictates otherwise. In addition, throughout the present specification, when a certain component is said to "include", it means that it may further include other components without excluding other components unless otherwise stated.

It is apparent to those skilled in the art that the present invention may be embodied in other specific forms without departing from the essential characteristics of the present invention.

Hereinafter, with reference to the accompanying drawings, a preferred embodiment will be described with reference to the accompanying drawings. In this process, the thickness of lines or the size of components shown in the drawings may be exaggerated for clarity and convenience of description. In addition, terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or operator. Therefore, definitions of these terms will have to be made based on the content throughout this specification.

In addition, the following examples do not limit the scope of the present invention, but are presented as examples only, and there may be various embodiments implemented through the technical idea.

1 is a flowchart illustrating a checkup order assignment method using an artificial intelligence learning model according to an embodiment of the present invention. FIG. 2 is a schematic configuration diagram illustrating a checkup order assignment system using an artificial intelligence learning model according to the embodiment of FIG. 1 .

Examination data collection step (S100)

In the examination data collection step (S100), the data collection unit 100 collects patient characteristic data according to the characteristics of the patient and examination time required for each examination item by the patient. Each step can proceed in the following process.

First, register patient information. A web-based or standalone application can display the patient's name, age, gender, and medical history. This information is then stored in a database such as MySQL or MongoDB.

The data collection unit 100 for learning the stored data automatically collects information about patients and examinations. This includes the use of information between logically connected databases and, if necessary, the collection of physical information such as the patient's height and weight through a separate sensor measurement device on site.

In the data collection unit 100, the collected data is processed and analyzed in a computer system to extract patient-related information. Through this process, data on the characteristics of each inspection item is separately stored in a database for learning.

In particular, the time required for a specific checkup item by a patient with the corresponding characteristic is actually measured and stored in a database for learning, and this collection process is continuously collected as the actual patient continues the checkup.

This collected data store is stored in a secure database such as a SQL or NoSQL database for future use. In order to prevent leakage of patient information, a method of maintaining security by separating the patient's personal information and actual collected information so that they cannot be checked in one database can be applied.

Model learning step (S200)

In the model learning step (S200), the model learning unit 200 learns the patient characteristic data and the required examination time, and creates a model capable of predicting the examination required time according to the patient characteristic data.

In the model learning step (S200), the model learning unit 200 learns patient characteristic data and the required examination time using various machine learning techniques, and creates a model capable of predicting the required examination time according to the patient characteristic data. From a computational point of view, the model training process may include the following steps:

Data pre-processing may be performed if necessary. Patient characteristic data and required examination time data collected through data collection may be pre-processed for the model learning process. This may include cleaning data, handling missing values, scaling or normalizing data, and more.

Regarding the selection of the model applied thereto, a machine learning algorithm suitable for the task of predicting the necessary test may be selected. Time according to patient characteristics. The choice of algorithm can be based on the nature of the data, the size of the data set and the complexity of the problem.

There are several machine learning model algorithms that can be applied to process patient data, depending on the nature of the data and the problem to be solved. The most commonly used algorithms are:

First, a linear regression algorithm may be applied. Linear regression is a supervised learning algorithm used to predict a continuous outcome variable based on a set of input features. It can be used to model the relationship between patient characteristics and examination duration.

A decision tree algorithm may be applied. A decision tree is a type of supervised learning algorithm used in classification and regression problems. Based on the input features, a tree-like decision structure can be created and used to model the relationship between patient characteristics and examination duration.

A random forest algorithm may be applied. It is a type of ensemble learning algorithm that combines multiple decision trees to improve the accuracy of a random forest model, and can be used to model the relationship between patient characteristics and test duration.

A gradient boosting algorithm may be applied. Gradient boosting is another type of ensemble learning algorithm that creates a set of weak models. And combine them to create a powerful model. This can be used to model the relationship between patient characteristics and examination duration.

In addition, neural network algorithms may be applied. Neural networks are the structure and function of the human brain. It can be used for a variety of tasks, including image and speech recognition, natural language processing, and prediction tasks.

Algorithms vary depending on the specific problem, available data, and available resources, so it is important to evaluate the performance of several algorithms and choose the one that best suits your specific use case.

The method used to perform learning is that the selected machine learning algorithm is trained on a prepared data set using techniques such as supervised or unsupervised learning. The algorithm uses input data and corresponding output data to learn the relationship between patient characteristics and test duration.

Evaluate the trained model using different performances. Metrics such as accuracy, precision, recall, and F1 score are used to determine a model's ability to generalize to new, never-before-seen data. At this time, the hyperparameters of the model can be fine-tuned for optimization.

Waiting number identification step (S300)

In the waiting number determination step (S300), the waiting number calculation unit 300 determines the number of waiting patients for each examination item. This process may include the following steps.

First, a search is performed on the data. Standby number calculation unit 300 retrieves data for patients and test items from the database. This is to first identify the status of potential waiting patients.

The retrieved data is processed using computer algorithms to determine the number of patients waiting for each test item. This may include calculating the number of patients scheduled for each test item and subtracting the number of patients who have already completed the test.

In this process, various methods for data visualization can be used. The number of waiting patients determined for each test item can be visualized using a data visualization tool. In particular, contents visualized in order to grasp the difference between the waiting situation on the computer and the actual situation of the patient waiting can be provided through a hospital manager's terminal or an installed monitor.

Air number calculation unit 300 is implemented using a combination of hardware and software components. Hardware components may include servers, storage devices, and other necessary equipment, while software components may include database management systems, programming languages, libraries, and visualization tools. It is also possible to host the database using a cloud service.

(c2) Application of beacons

The waiting number calculation unit 300 includes a beacon unit 310, and uses Bluetooth technology to determine the number of waiting people for each examination item in conjunction with the patient's smartphone. This process may include the following steps:

First, the beacon unit 310 is placed in advance. Beacons can be installed in hospital examination rooms and waiting rooms, and emit Bluetooth signals that can be received by patients' smartphones.

Patients need to download and install a clinic or hospital-specific app on their smartphone. The app can detect beacon signals and locate a patient within a clinic or hospital. If possible, a method that does not use an app can be used to determine the location of a specific patient in a limited manner within the hospital.

In addition, when a patient walks around a clinic or hospital, the patient's smartphone detects the beacon signal and transmits location information. The waiting number calculation unit 300 uses this information to determine the number of patients waiting for each examination item in real time.

As a result, the waiting number calculation unit 300 uses the beacon unit 310 combined with the patient's smart phone to identify people waiting for each examination item. The beacon unit 310 detects the location of a patient's smartphone in a clinic or hospital using Bluetooth technology and sends this information to the waiting number calculating unit 300 to calculate the number of waiting patients for each test item in real time.

Screening item assignment step (S400)

In the examination item allocation step (S400), the examination order allocation unit 400 allocates the examination items optimized for the examination patient in real time. Conditions that can be applied in this case are the patient's characteristic data and the examination time of the corresponding examination item expected accordingly.

Based on the elements of these two groups, it is possible to provide patients and management with a sequence of patient checkups that takes the least amount of time.

In particular, when a patient first starts a checkup, these optimized checkup items may be provided in order from the beginning, but a timeline for each checkup item may be simulated using an estimated checkup time using patient characteristic data, and at the same time, a timeline for each patient may be simulated.

Accordingly, a more optimized examination operation can be performed by providing the examination order to the patient through the entire predicted schedule.

In particular, this checkup order enables more efficient progress by modifying and providing the next checkup item based on the patient in accordance with the changing situation in real time.

Therefore, when a patient completes an examination, the true order allocation unit 400 updates the examination order in real time so that the allocation is always optimized. A detailed proceeding method for this will be described along with other drawings.

(e) time information providing step (S500)

In the time information providing step (S500), the time calculation unit 500 provides the waiting time and final completion time to the patient according to the patient characteristic data and the number of waiting patients.

The time calculation unit 500 may send notifications and warnings, such as an expected waiting time for a specific examination item, to patients and staff in real time. or when it is the patient's turn to be examined. This can improve the overall patient experience and help staff manage patient flow more effectively.

In particular, data collected by the beacon unit 310 may be integrated with other systems. An electronic medical record (EMR) or hospital information system (HIS) can be used to improve overall patient care by providing more comprehensive access to patient data.

In particular, an app installed to communicate with the beacon unit 310 can be applied as a means to more efficiently provide information to the patient. Through this, the patient is notified of the next checkup location, and when a change occurs, it is changed to the next location, and the total time reduced is simultaneously informed, thereby providing information that is helpful to the patient.

How to assign screening order

3 is a flowchart illustrating a step of allocating checkup items in a checkup order assignment method using an artificial intelligence learning model according to another embodiment of the present invention. 4A to 4B are data sets for explaining a checkup order allocation method using the artificial intelligence learning model of FIG. 1 . 5A to 5B are timeline sets for explaining a checkup order assignment method using the artificial intelligence learning model of FIG. 1 .

The examination item assignment step (S400) includes a time line configuration step (S410), an assignment exchange step (S420), a repeat execution step (S430), and a checkup item assignment step (S440).

First, in the timeline configuration step (S410), a timeline including a start time, an end time, and a waiting time for each examination item is predicted and configured for all of the plurality of patients.

In the timeline configuration step, a schedule table including start time, end time, and waiting time for each examination item scheduled for treatment for each patient is prepared. In particular, the checkup time for a specific item of the patient can be matched with the patient's characteristic data using the generated model to obtain a result value, so the final timeline can be simulated based on this.

Referring first to FIG. 4A , in a state in which characteristic data for each patient is input, an expected examination time for each patient can be estimated as shown in the table of FIG. 4A . However, this is not a definitive checkup time because various variables may occur during the checkup and influencing factors outside the patient's characteristic data are not considered. However, simulating this is used to select the optimal method based on the provided data.

Instead, based on these data, a schematic patient timeline can be established, as shown in FIG. 5A.

FIG. 5A is a table showing waiting times and checkup statuses in checkup items for each category for patients C and E.

Looking at the timeline for each patient, the end time can be predicted using information on the checkup sequence for each checkup item, the start time for each individual checkup item, and the time required for checkup for each patient's individual checkup item, etc., and thus the overall waiting time. Finally, the patient can specify the time at which the examination is completed.

In the allocation exchange step (S420), when the first examination item of a patient and the second examination item of another patient are exchanged, if the final completion time of the patient or the other patient decreases, the first examination item of the patient and the second examination item of the other patient are exchanged.

The table of FIG. 4B shows the result of shifting the times in examination item 1 of patient C and patient E to each other. Looking at this as a timeline for each patient, it is the same as the timeline set of FIG. 5B.

Referring again to FIG. 5B , a total of 18 minutes is required for the final checkup time of patient C, as checkup item 3 is shifted and checkup item 2, checkup item 1, and checkup item 3 are finally performed. The reason why checkup item 3 is shifted is that the completion time of checkup item 1, which was physically performed, ends later than the scheduled start time of checkup item 3.

When determining the timeline for each patient and each checkup item, if the physical time for each patient overlaps, the waiting time for each checkup item is additionally assigned and the checkup start time of the patient is shifted and calculated.

As a result of switching the checkup order of patient C and patient E in checkup item 1, the total checkup time of patient C was shortened by 1 minute. In addition, since the total checkup time for patient E is 17 minutes in total, patient E's schedule is not affected, and patient C's waiting time is reduced by that much.

The exchange of examination items for each patient is repeated in real time. In the repeating step (S430), when one checkup item for the random patient ends differently from the estimated time required for the checkup, the exchange assignment step (S420) is repeated in real time based on the checkup item for the random patient.

In particular, since the expected checkup time can increase or decrease depending on the actual checkup situation, as soon as these variables occur, the entire schedule is adjusted, conditions that can be readjusted based on the changed environment are explored, and when the conditions are satisfied, the patient's checkup items are exchanged with each other.

If this is implemented as a program, processing and confirmation of hundreds of cases per minute is possible, so that immediately changing conditions can be reflected in real time, and therefore, patients and hospitals can proceed with the examination while taking the optimal time.

Examination Item Assignment Step (S440) The next examination item optimized for the end point of the patient who has completed the one examination item is assigned. Since there is a possibility that the patient's checkup order can be changed before the patient completes one checkup, accordingly, the exchange assignment step (S420) is checked and performed at high speed by a computer process before the checkup item is assigned (S440).

Accordingly, as a result, the patient can efficiently perform examination while consuming a minimum of time.

The present invention is not limited by the foregoing embodiments and accompanying drawings. It will be clear to those skilled in the art that the components according to the present invention can be substituted, modified, and changed without departing from the technical spirit of the present invention.

100: data collection unit
200: model learning unit
300: waiting number calculation unit
310: beacon unit
400: examination order allocation unit
500: time calculation unit

Claims (10)

In the data collection unit 100, a checkup data collection step (S100) of collecting patient characteristic data according to the characteristics of the patient and a checkup time required for each checkup item by the patient;
Learning the examination time model in the model learning unit 200, learning the patient characteristic data and the required examination time, and generating a model capable of predicting the examination required time according to the patient characteristic data (S200);
In the waiting number calculation unit 300, a waiting number determination step of determining the number of waiting patients for each current examination item (S300);
a checkup item allocating step (S400) of allocating checkup items to the checkup patient in real time in the checkup order assigning unit 400; and
and a time information providing step (S500) of providing waiting time and final completion time to the patient according to the patient characteristic data and the number of waiting patients in the time calculation unit 500.
According to claim 1,
The waiting number calculation unit 300 includes a beacon unit 310,
The checkup order assignment method, characterized in that the beacon unit 310 identifies the number of people waiting for each item in association with the patient's smartphone.
According to claim 1,
The examination item allocation step (S400) uses the examination order allocation unit 400,
a time line configuration step of predicting and constructing a timeline including start time, end time, and waiting time for each of the examination items for all of the plurality of patients (S410);
When the first examination item of any patient and the second examination item of any other patient are exchanged, if the final completion time of the patient or any other patient decreases, an allocation exchange step of exchanging the first examination item of the arbitrary patient and the second examination item of the other arbitrary patient with each other (S420);
If one checkup item for the random patient ends differently from the estimated time required for the checkup, repeating the exchange allocation step (S420) in real time based on the checkup item for the random patient (S430); and
and a checkup item allocating step (S440) of allocating the next checkup item at the end point of the patient who has completed the one checkup item.
According to claim 3,
In the timeline construction step (S410),
and shifting the start time of the one checkup item to an end time of another checkup item when the end time of another checkup item is later than the start time of the checkup item of the patient.
According to claim 3,
In the timeline construction step (S410),
and predicting and constructing a timeline including the examination start time, end time, and waiting time for all patients with respect to a plurality of all examination items.
a data collection unit 100 that collects patient characteristic data according to the characteristics of the patient and examination time required for each examination item by the patient;
A model learning unit 200 for learning the patient characteristic data and the required examination time and generating a model capable of predicting the examination required time according to the patient characteristic data;
Waiting number calculation unit 300 for determining the number of waiting patients for each current examination item;
an examination order assigning unit 400 for allocating examination items to the examination patients in real time; and
and a time calculation unit 500 configured to provide a waiting time and a final completion time to the patient according to the patient characteristic data and the number of waiting patients.
According to claim 6,
The waiting number calculation unit 300 includes a beacon unit 310,
The checkup order assignment system, characterized in that the beacon unit 310 identifies the number of people waiting for each item in association with the patient's smartphone.
According to claim 6,
The checkup order assignment unit 400
a time line configuration unit that predicts and constructs a time line including a start time, an end time, and a waiting time for each of the checkup items for all of the plurality of patients;
an assignment exchange unit that exchanges the first examination item of any patient and the second examination item of any other patient, if the final completion time of the patient or any other patient decreases when the first examination item of any patient and the second examination item of any other patient are exchanged with each other;
a repetition implementation unit repeating the exchange assignment step (S420) in real time based on the examination item for the random patient when one examination item for the random patient ends differently from the estimated time required for the examination; and
and a checkup item assigning unit for allocating the next checkup item at an end point of the patient who has completed the one checkup item.
According to claim 6,
The timeline component,
When the end time of another checkup item is later than the start time of the patient's individual checkup item, the start time of the one checkup item is shifted to after the end time of the other checkup item.
According to claim 6,
The timeline component,
The examination order assignment system, characterized in that for predicting and configuring a timeline including the examination start time, the examination end time, and the waiting time for all patients for all examination items.