KR20230120163A - Method for providing information of intradialytic hypotension and device using the same - Google Patents

Abstract

The present invention is a method for providing information on hypotension during hemodialysis implemented by a processor, comprising the steps of receiving hemodialysis machine data and vital sign data obtainable during hemodialysis from an individual, and the hemodialysis machine data and vital sign data as inputs. predicting hypotension during dialysis based on the received hemodialysis machine data and vital sign data, using a prediction model learned to predict hypotension during dialysis within a predetermined time period, wherein the hemodialysis machine data is received from the hemodialysis machine. Provided is a method for providing information on hypotension during dialysis, which is time-series data obtained in units of minutes, and a device using the same.

Description

Method for providing information on hypotension during dialysis and device for providing information on hypotension during dialysis using the same

The present invention relates to a method for providing information on hypotension during dialysis and a device for providing information on hypotension during dialysis using the same.

Intradialytic hypotension (IDH) is one of the most common complications of chronic hemodialysis. At this time, diabetes, cardiovascular disease, autonomic nervous system dysfunction, nutritional status, old age, anemia, high weight gain between dialysis, and the like are known risk factors associated with hypotension during dialysis. However, most of these risk factors are difficult to correct immediately during hemodialysis. Accordingly, when hypotension occurs during dialysis, treatment such as temporarily stopping hemodialysis or lowering the ultrafiltration rate is performed.

On the other hand, it may be helpful to measure blood pressure more frequently in order to detect hypotension at an early stage during dialysis, but it is impossible to continuously measure blood pressure during dialysis due to the characteristics of a non-invasive blood pressure measurement method. Furthermore, because hypotension during dialysis has various causes, it can be difficult to predict with classical statistical models.

Accordingly, other non-invasive methods capable of predicting hypotension during dialysis in advance may be required.

Therefore, there is a continuous demand for the development of a method for providing reliable information related to hypotension during dialysis.

The background description of the invention has been prepared to facilitate understanding of the present invention. It should not be construed as an admission that matters described in the background art of the invention exist as prior art.

On the other hand, in order to solve the above-described problems, a study on predicting hypotension during dialysis of a hemodialysis patient using machine learning or deep learning capable of processing multi-dimensional data has been introduced.

However, since the corresponding model requires various clinical information about the individual, there may be a risk of personal information infringement.

The inventors of the present invention, as a way to overcome the above-mentioned limitations, blood flow rate, ultrafiltration rate, dialysis fluid flow rate, arterial vascular pressure, venous vascular pressure, transmembrane pressure, temperature, and various parameters including bicarbonate and sodium levels in the dialysis fluid during a hemodialysis session. It was noted that it is possible to obtain measured values.

More specifically, the inventors of the present invention have noted that the above-described measurement values are easily obtainable from hemodialysis machines, as they are obtainable from all hemodialysis machines of various manufacturers.

In particular, the inventors of the present invention intend to build a model capable of predicting the risk of developing hypotension during dialysis by using data generated in a hemodialysis machine as learning data, excluding data based on personal information such as electronic medical record data. did

As a result, the inventors of the present invention have developed a new system for providing information on hypotension during dialysis based on a predictive model.

On the other hand, the inventors of the present invention collect real-time data in units of minutes during which the dialysis machine is monitored for a predetermined time, and then use this as learning data to predict hypotension risk after a specific time (eg, 10 minutes later) We tried to build a predictive model to do this.

As a result, the inventors of the present invention can quickly predict hypotension during hemodialysis by the predictive model, so medical staff's treatment for hypotension, such as Trendelenburg position, isotonic solution administration, and ultrafiltration rate reduction, is effectively performed. could be expected to be.

Moreover, the inventors of the present invention tried to build a predictive model to predict hypotension during hemodialysis in various forms.

More specifically, the inventors of the present invention, "Fall 20" systolic blood pressure is reduced by 20 mmHg or more compared to the initial blood pressure, which is defined as Nadir systolic blood pressure is less than 90 mmHg, and and/or could be constructed to predict hypotension during hemodialysis in three forms of “Fall20 or MAP10” in which mean arterial pressure decreased by more than 10 mmHg.

Accordingly, the inventors of the present invention could expect that by providing an information providing system based on a predictive model learned to predict the occurrence of hypotension during various types of dialysis, it could contribute to the application of individualized treatment by medical staff.

That is, by providing a new information providing system, the inventors of the present invention expected that it was possible to predict low blood pressure during dialysis regardless of the skill level of medical staff and to provide highly reliable information therefor.

Therefore, the problem to be solved by the present invention is to provide a method for providing information on hypotension during dialysis configured to predict the occurrence of hypotension during dialysis based on clinical data obtained from an individual using a predictive model, and to provide a device using the same will be.

The tasks of the present invention are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the following description.

In order to solve the above problems, a method for providing information on hypotension during dialysis according to an embodiment of the present invention is provided. The information providing method is a method of providing information on hypotension during hemodialysis implemented by a processor, comprising the steps of receiving hemodialysis machine data and vital sign data obtainable during hemodialysis from an individual, and the hemodialysis machine data and vital sign data and predicting hypotension during dialysis based on the received hemodialysis machine data and vital sign data, using a predictive model learned to predict hypotension during dialysis within a predetermined time as an input.

In this case, the hemodialysis machine data is time-series data obtained from the hemodialysis machine in units of minutes.

According to a feature of the present invention, the hemodialysis machine data includes total dialysis time, arterial line pressure (AP), venous line pressure (VP), blood flow rate, obtained from the hemodialysis machine in minutes over a predetermined time period. It may be at least one of a blood flow rate, a dialysate flow rate, an ultrafiltration rate, a total ultrafiltration volume, a dialysate temperature, and a sodium level.

According to another feature of the present invention, the hemodialysis machine data may be blood flow rate, dialysate sodium level and arterial pressure.

According to another feature of the present invention, the vital sign data may be at least one of systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and pulse rate. there is.

According to another feature of the invention, the vital sign data may be systolic blood pressure and mean arterial pressure.

According to another feature of the present invention, the vital sign data may be time-series data obtained from an individual on an hourly basis.

According to another feature of the present invention, the predetermined time may be 1 minute to 1 hour.

According to another feature of the present invention, the predicting step may include at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' based on the received hemodialysis machine data and vital sign data using a predictive model. of predicting hypotension during dialysis.

At this time, Nadir90 is defined as a state in which Nadir systolic blood pressure is less than 90 mmHg, Fall 20 is defined as a state in which systolic blood pressure is 20 mmHg or more compared to the initial blood pressure, and Fall 20 or MAP10 is a decrease in systolic blood pressure of 20 mmHg or more and / Alternatively, it can be defined as a decrease in mean arterial pressure of 10 mmHg or more.

According to another feature of the present invention, the predictive model may include a convolutional neural network (CNN) module for inputting hemodialysis machine data and the vital sign data, and a feedforward neural network (FNN) module for outputting hypotension during dialysis. can

According to another feature of the present invention, the CNN module includes an input layer, a computation block consisting of a plurality of computation layers, and an adaptive average pooling layer, and the FNN module includes an active function layer, a dropout layer, and It may contain an output layer.

According to another feature of the present invention, the plurality of computation layers may include at least two layers of a 1D convolution layer, a 1D average pooling layer, a batch normalization layer, and an activation function layer.

In order to solve the above problems, a device for providing information on hypotension during dialysis according to an embodiment of the present invention is provided. The device includes a communication unit configured to receive hemodialysis machine data and vital sign data obtainable during hemodialysis from an individual, and a processor functionally connected to the communication unit.

At this time, the processor uses a prediction model learned to predict hypotension during dialysis within a predetermined time by taking the hemodialysis machine data and vital sign data as inputs, and hypotension during dialysis based on the received hemodialysis machine data and vital sign data. and the hemodialysis machine data is time-series data obtained from the hemodialysis machine in minutes.

Other embodiment specifics are included in the detailed description and drawings.

The present invention has an effect of providing information related to hypotension during dialysis by providing an information providing system based on an artificial neural network configured to predict hypotension during dialysis using clinical data excluding personal information.

In particular, since the present invention provides a predictive model capable of predicting hypotension after a predetermined time (eg, 10 minutes) during dialysis, medical personnel can quickly perform various treatments.

More specifically, the present invention can contribute to the effective treatment of hypotension by medical staff, such as the Trendelenburg posture, isotonic solution administration, and ultrafiltration rate reduction, since hypotension can be quickly predicted during hemodialysis by a predictive model.

Furthermore, the present invention may provide a predictive model-based information providing system configured to predict hypotension during various types of hemodialysis.

More specifically, the present invention can predict hypotension during hemodialysis in three types of “Nadr90” “Fall20” and “Fall20 or MAP10”.

That is, the present invention provides an information providing system based on a predictive model learned to predict the occurrence of hypotension during various types of dialysis, thereby contributing to enabling medical personnel to easily select individualized treatments for each individual.

Accordingly, the present invention provides a new information providing system, so that it is possible to predict low blood pressure during dialysis regardless of the skill level of a medical staff, and to provide highly reliable information therefor.

Effects according to the present invention are not limited by the contents exemplified above, and more various effects are included in the present specification.

1 illustrates a system for providing information on hypotension during dialysis using a device for providing information on hypotension during dialysis according to an embodiment of the present invention.
2A is a block diagram showing the configuration of a medical staff device according to an embodiment of the present invention.
2B is a block diagram showing the configuration of a server for providing information according to an embodiment of the present invention.
3 illustrates a procedure of a method for providing information on hypotension during dialysis according to an embodiment of the present invention.
4 illustratively illustrates a procedure of a method for providing information on hypotension during dialysis according to an embodiment of the present invention.
FIG. 5 illustratively illustrates the structure of a predictive model used in a method for providing information on hypotension during dialysis according to an embodiment of the present invention.
6a to 6f illustrate learning data and verification results of a predictive model in a method for providing information on hypotension during dialysis according to various embodiments of the present invention.
7A to 7C show main data for prediction of hypotension during dialysis by a predictive model in a method for providing information on hypotension during dialysis according to various embodiments of the present invention.
8A to 8C illustrate main data for prediction of hypotension during dialysis by a predictive model in a method for providing information on hypotension during dialysis according to various embodiments of the present invention.

The advantages of the invention, and how to achieve them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be implemented in various different forms, only these embodiments make the disclosure of the present invention complete, and common knowledge in the art to which the present invention belongs. It is provided to fully inform the holder of the scope of the invention, and the present invention is only defined by the scope of the claims.

The shapes, sizes, ratios, angles, numbers, etc. disclosed in the drawings for explaining the embodiments of the present invention are illustrative, so the present invention is not limited to the details shown. In addition, in describing the present invention, if it is determined that a detailed description of related known technologies may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. When 'includes', 'has', 'consists of', etc. mentioned in this specification is used, other parts may be added unless 'only' is used. In the case where a component is expressed in the singular, the case including the plural is included unless otherwise explicitly stated.

In interpreting the components, even if there is no separate explicit description, it is interpreted as including the error range.

Each feature of the various embodiments of the present invention can be partially or entirely combined or combined with each other, and as those skilled in the art can fully understand, various interlocking and driving operations are possible, and each embodiment can be implemented independently of each other. It may be possible to implement together in an association relationship.

For clarity of interpretation of this specification, terms used in this specification will be defined below.

As used herein, the term "subject" may refer to any subject who wishes to receive information on hypotension during dialysis. For example, the subject may be any mammal except for humans, but is not limited thereto.

As used herein, the term “hemodialysis machine data” may refer to all data obtainable from a hemodialysis machine.

According to a feature of the present invention, the hemodialysis machine data includes total dialysis time, arterial line pressure (AP), venous line pressure (VP), blood flow rate, obtained from the hemodialysis machine for a predetermined period of time. ), dialysate flow rate, ultrafiltration rate, total ultrafiltration volume, dialysate temperature, and dialysate sodium level, but is not limited thereto.

Preferably, the hemodialysis machine data in the present specification may be a blood flow rate in units of minutes, a dialysate sodium level, and an arterial pressure obtained for a predetermined period of time. For example, the hemodialysis machine data may be blood flow rate, dialysate sodium level, and arterial pressure acquired every minute for 30 minutes. Accordingly, hemodialysis machine data may exist in a plurality of segments. However, it is not limited thereto.

Meanwhile, the above-described hemodialysis machine data may be obtained from any hemodialysis machine regardless of manufacturer and furthermore, without limitation of personal information.

As used herein, the term “vital sign data” may refer to all data related to blood pressure, pulse rate, respiratory rate, and body temperature obtainable from an individual.

According to a feature of the present invention, vital sign data includes systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP) and pulse rate obtained on an hourly basis from a subject. may be at least one of However, it is not limited thereto.

Preferably, the vital sign data herein may be systolic blood pressure and mean arterial pressure obtained on an hourly basis from an individual. For example, the vital sign data may be systolic blood pressure and mean arterial pressure obtained on an hourly basis. However, it is not limited thereto.

As used herein, the term "intradialytic hypotension (IDH)" may refer to hypotension occurring in hemodialysis.

According to a feature of the present invention, hypotension during dialysis may be at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10'.

In this case, Nadir90 may be defined as a state in which the Nadir systolic blood pressure is less than 90 mmHg, and Fall 20 may be defined as a state in which the systolic blood pressure is 20 mmHg or more compared to the initial blood pressure. Furthermore, Fall20 or MAP10 may be defined as a state in which systolic blood pressure decreases by 20 mmHg or more and/or mean arterial pressure decreases by 10 mmHg or more.

As used herein, the term "prediction model" refers to the occurrence of hypotension within a predetermined time by using hemodialysis machine data and vital sign data excluding personal information-based data among clinical data obtainable from an individual undergoing dialysis as inputs. It may be a model configured to output

According to a feature of the present invention, the predictive model is trained to predict the onset of three classes of hypotension during dialysis, 'Nadir90', 'Fall20', and 'Fall20 or MAP10', based on hemodialysis machine data and vital sign data. may be a model.

According to another feature of the present invention, the predictive model may include a convolutional neural network (CNN) module for inputting hemodialysis machine data and the vital sign data, and a feedforward neural network (FNN) module for outputting hypotension during dialysis.

At this time, the CNN module includes an input layer, a computation block composed of a plurality of computation layers, and an adaptive average pooling layer, and the FNN module may include an active function layer, a dropout layer, and an output layer. .

According to another feature of the present invention, the plurality of computation layers may include at least two layers of a 1D convolution layer, a 1D average pooling layer, a batch normalization layer, and an activation function layer.

According to another feature of the present invention, in constructing a predictive model, 10 time-invariant variables are processed as a fixed-valued one-dimensional image, and time-varying so that correlations between features are captured. Time-varying variables and time-invariant variables can be spliced into a one-dimensional image with 30 pixels and 20 channels.

However, the model construction process is not limited thereto.

On the other hand, the predictive model is a logistic regression (LR) based model, a random forest (RF) based model, an extreme gradient boosting (XGB) based model, and a deep learning model ; DLM). However, it is not limited thereto.

For example, prediction models include U-net VGG net, DenseNet, FCN (Fully Convolutional Network) having an encoder-decoder structure, SegNet, DeconvNet, DNN (deep neural network) such as DeepLAB V3+, SqueezeNet, Alexnet, and ResNet18. , MobileNet-v2, GoogLeNet, Resnet-v2, Resnet50, RetinaNet, Resnet101, and Inception-v3. Furthermore, the predictive model may be an ensemble model based on at least two algorithm models among the aforementioned algorithms.

Hereinafter, with reference to FIGS. 1 and 2a and 2b, a system for providing information on hypotension during dialysis using a device for providing information on hypotension during dialysis and a device for providing information on hypotension during dialysis according to an embodiment of the present invention are provided. Explain.

1 illustrates a system for providing information on hypotension during dialysis using a device for providing information on hypotension during dialysis according to an embodiment of the present invention. FIG. 2A illustratively illustrates the configuration of a medical staff device receiving information on hypotension during dialysis according to an embodiment of the present invention. Figure 2b shows the configuration of a device for providing information on hypotension during dialysis according to an embodiment of the present invention by way of example.

First, referring to FIG. 1 , an information providing system 1000 may be a system configured to provide information related to hypotension during dialysis based on hemodialysis machine data and vital sign data of an individual. At this time, the information providing system 1000 includes the medical staff device 100 for receiving information related to hypotension during dialysis, the hemodialysis machine 200 for providing hemodialysis machine data during hemodialysis, and vitality for providing vital sign data for an individual. It may be configured with a symptom measuring device (not shown) and an information providing server 300 that generates information about low blood pressure during dialysis based on the received hemodialysis machine data and vital sign data.

First, next, the medical staff device 100 is an electronic device that provides a user interface for displaying information related to hypotension during dialysis, and uses at least one of a smart phone, a tablet PC (Personal Computer), a laptop computer, and/or a PC. can include

The medical staff device 100 may receive a prediction result associated with hypotension during dialysis of an individual from the server 300 for providing information, and display the received result through a display unit to be described later.

The server 300 for providing information is a general purpose server that performs various calculations for determining information related to hypotension during dialysis based on the hemodialysis machine data and vital sign data provided from the hemodialysis machine 200 and a vital sign measurement device (not shown). computers, laptops, and/or data servers, and the like. In this case, the information providing server 300 may be a device for accessing a web server providing web pages or a mobile web server providing mobile web sites, but is not limited thereto.

More specifically, the server 300 for providing information receives hemodialysis machine data and vital sign data from the hemodialysis machine 200 and vital sign measuring device (not shown), and based on the received hemodialysis machine data and vital sign data It is possible to provide information related to hypotension during dialysis by determining whether hypotension occurs within a predetermined time during dialysis. In this case, information related to low blood pressure during dialysis may be predicted from hemodialysis machine data and vital sign data using the information providing server 300 and a predictive model.

The server 300 for providing information may provide a prediction result associated with low blood pressure during dialysis to the medical device 100 .

The information provided from the server 300 for providing information in this way may be provided as a web page through a web browser installed in the medical device 100, or may be provided in the form of an application or program. In various embodiments, this data may be provided in a form incorporated into the platform in a client-server environment.

Next, with reference to FIGS. 2A and 2B, components of the server 300 for providing information according to the present invention will be described in detail.

First, referring to FIG. 2A , the medical device 100 may include a memory interface 110 , one or more processors 120 and a peripheral interface 130 . The various components within clinician device 100 may be connected by one or more communication buses or signal lines.

The memory interface 110 is connected to the memory 150 and can transfer various data to the processor 120 . Here, the memory 150 is a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg SD or XD memory, etc.), RAM, SRAM, ROM, EEPROM, PROM, network storage storage, cloud , It may include at least one type of storage medium among blockchain databases.

In various embodiments, memory 150 includes operating system 151 , communication module 152 , graphical user interface module (GUI) 153 , sensor processing module 154 , telephony module 155 , and application module 156 . ) At least one or more of them may be stored. Specifically, the operating system 151 can include instructions for handling basic system services and instructions for performing hardware tasks. The communication module 152 can communicate with at least one of another one or more devices, computers, and servers. A graphical user interface module (GUI) 153 can process a graphical user interface. Sensor processing module 154 can process sensor-related functions (eg, process voice input received using one or more microphones 192). The telephony module 155 can process telephony-related functions. Application module 156 can perform various functions of a user application, such as electronic messaging, web browsing, media processing, navigation, imaging, and other processing functions. In addition, the medical device 100 may store one or more software applications 156 - 1 and 156 - 2 (eg, information providing applications) associated with a certain type of service in the memory 150 .

In various embodiments, memory 150 can store digital assistant client module 157 (hereafter referred to as DA client module), thereby storing instructions and various user data 158 for performing client-side functions of the digital assistant. (eg, user-customized vocabulary data, preference data, and other data such as the user's electronic address book, to-do list, shopping list, etc.).

Meanwhile, the DA client module 157 receives a user's voice input, text input, touch input, and/or gesture through various user interfaces (eg, the I/O subsystem 140) provided in the medical device 100. input can be obtained.

In addition, the DA client module 157 can output data in audio-visual and tactile forms. For example, the DA client module 157 may output data consisting of a combination of at least two of voice, sound, notification, text message, menu, graphic, video, animation, and vibration. In addition, the DA client module 157 can communicate with a digital assistant server (not shown) using the communication subsystem 180 .

In various embodiments, DA client module 157 can collect additional information about the surrounding environment of clinician device 100 from various sensors, subsystems, and peripheral devices to construct a context associated with user input. . For example, the DA client module 157 can provide contextual information along with user input to the digital assistant server to infer the user's intent. Here, the situational information that may accompany the user input may include sensor information, eg, lighting, ambient noise, ambient temperature, image of the surrounding environment, video, and the like. For another example, contextual information can include the physical state of the clinician device 100 (eg, device orientation, device location, device temperature, power level, speed, acceleration, motion patterns, cellular signal strength, etc.) . For another example, contextual information may include information related to the state of the software of the clinician device 100 (e.g., processes running on the clinician device 100, installed programs, past and current network activity, background services, error logs, resource usage, etc.).

In various embodiments, the memory 150 may include added or deleted commands, and the medical device 100 may also include additional components other than those shown in FIG. 2A or may exclude some components.

The processor 120 can control the overall operation of the medical device 100, and executes various commands for implementing an interface that provides information related to hypotension during dialysis by driving an application or program stored in the memory 150. can

The processor 120 may correspond to a computing device such as a central processing unit (CPU) or an application processor (AP). In addition, the processor 120 may be implemented in the form of an integrated chip (IC) such as a System on Chip (SoC) in which various computing devices such as a Neural Processing Unit (NPU) are integrated.

The peripheral interface 130 may be connected to various sensors, subsystems, and peripheral devices to provide data so that the medical device 100 can perform various functions. Here, the fact that the medical device 100 performs a certain function can be understood as being performed by the processor 120 .

Peripheral interface 130 can receive data from motion sensor 160, light sensor (light sensor) 161, and proximity sensor 162, through which medical device 100 can determine orientation, light, and proximity. sensing function, etc. For another example, peripheral interface 130 can receive data from other sensors 163 (positioning system-GPS receiver, temperature sensor, biometric sensor), through which medical device 100 can receive data from other sensors. (163) and related functions can be performed.

In various embodiments, clinician device 100 can include a camera subsystem 170 coupled with peripheral interface 130 and an optical sensor 171 coupled thereto, through which clinician device 100 can take pictures and video Various shooting functions such as clip recording can be performed.

In various embodiments, clinician device 100 can include a communication subsystem 180 coupled with peripheral interface 130 . The communication subsystem 180 is composed of one or more wired/wireless networks, and may include various communication ports, radio frequency transceivers, and optical transceivers.

In various embodiments, clinician device 100 includes an audio subsystem 190 coupled to peripheral interface 130, which audio subsystem 190 includes one or more speakers 191 and one or more microphones 192. By inclusion, clinician device 100 can perform voice-activated functions, such as voice recognition, voice replication, digital recording, and telephony functions.

In various embodiments, clinician device 100 can include I/O subsystem 140 coupled with peripheral interface 130 . For example, the I/O subsystem 140 can control the touch screen 143 included in the medical practitioner device 100 via the touch screen controller 141 . As an example, the touch screen controller 141 uses any one of a plurality of touch sensing technologies such as capacitive, resistive, infrared, surface acoustic wave technology, proximity sensor array, and the like to detect the user's touch and movement or touch. and cessation of movement. For another example, I/O subsystem 140 can control other input/control devices 144 included in clinician device 100 via other input controller(s) 142 . As an example, other input controller(s) 142 may control one or more buttons, rocker switches, thumb-wheels, infrared ports, USB ports, and pointer devices such as styluses and the like.

Next, referring to FIG. 2B , the information providing server 300 may include a communication interface 310, a memory 320, an I/O interface 330, and a processor 340, each component being one They can communicate with each other through the above communication buses or signal lines.

The communication interface 310 may be connected to the medical device 100, the hemodialysis machine 200, and vital sign measurement devices (not shown) through a wired/wireless communication network to exchange data. For example, the communication interface 310 can receive hemodialysis machine data and vital sign data from the hemodialysis machine 200 and a vital sign measuring device (not shown), and transmits information associated with the determined low blood pressure during hemodialysis to a medical staff device ( 100) can be sent.

On the other hand, the communication interface 310 enabling transmission and reception of such data includes a communication pod 311 and a wireless circuit 312, where the wired communication port 311 is one or more wired interfaces, for example, Ethernet, This may include Universal Serial Bus (USB), FireWire, and the like. In addition, the wireless circuitry 312 can transmit and receive data to and from external devices through RF signals or optical signals. In addition, wireless communication may use at least one of a plurality of communication standards, protocols and technologies, such as GSM, EDGE, CDMA, TDMA, Bluetooth, Wi-Fi, VoIP, Wi-MAX, or any other suitable communication protocol.

The memory 320 can store various data used in the server 300 for providing information. For example, memory 320 can store hemodialysis machine data and vital sign data, or a predictive model learned to predict hypotension on dialysis based on hemodialysis machine data and vital sign data.

In various embodiments, memory 320 may include volatile or non-volatile recording media capable of storing various data, commands, and information. For example, the memory 320 may be a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg SD or XD memory, etc.), RAM, SRAM, ROM, EEPROM, PROM, network storage storage. , Cloud, and a blockchain database may include at least one type of storage medium.

In various embodiments, memory 320 can store a configuration of at least one of operating system 321 , communication module 322 , user interface module 323 and one or more applications 324 .

Operating system 321 (e.g., an embedded operating system such as LINUX, UNIX, MAC OS, WINDOWS, VxWorks, etc.) controls and manages general system tasks (e.g., memory management, storage device control, power management, etc.) It may include various software components and drivers for doing so, and may support communication between various hardware, firmware, and software components.

The communication module 323 can support communication with other devices via the communication interface 310 . The communication module 320 may include various software components for processing data received by the wired communication port 311 or the wireless circuitry 312 of the communication interface 310 .

The user interface module 323 can receive a user's request or input from a keyboard, touch screen, microphone, etc. through the I/O interface 330 and provide a user interface on a display.

Applications 324 can include programs or modules configured to be executed by one or more processors 330 . Here, an application for providing information related to hypotension during dialysis may be implemented on a server farm.

The I/O interface 330 may connect at least one of an input/output device (not shown) of the information providing server 300, for example, a display, a keyboard, a touch screen, and a microphone, to the user interface module 323. The I/O interface 330 together with the user interface module 323 can receive user input (eg, voice input, keyboard input, touch input, etc.) and process commands according to the received input.

The processor 340 is connected to the communication interface 310, the memory 320, and the I/O interface 330 to control the overall operation of the server 300 for providing information, and applications stored in the memory 320 or Various commands for providing information can be executed through the program.

The processor 340 may correspond to a computing device such as a central processing unit (CPU) or an application processor (AP). In addition, the processor 340 may be implemented in the form of an integrated chip (IC) such as a System on Chip (SoC) in which various computing devices are integrated. Alternatively, the processor 340 may include a module for calculating an artificial neural network model, such as a Neural Processing Unit (NPU).

In various embodiments, processor 340 uses predictive models to determine on-dialysis hypotension, particularly at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' on-dialysis based on the hemodialysis machine data and vital sign data. It can be configured to predict whether hypotension will occur.

Hereinafter, a method for providing information on low blood pressure during dialysis according to an embodiment of the present invention will be described in detail with reference to FIGS. 3 and 4 . 3 illustrates a procedure of a method for providing information on hypotension during dialysis according to an embodiment of the present invention. 4 illustratively illustrates a procedure of a method for providing information on hypotension during dialysis according to an embodiment of the present invention.

First, referring to FIG. 3, an information providing procedure according to an embodiment of the present invention is as follows. First, hemodialysis machine data and vital sign data for an individual are received (S310). Then, whether hypotension occurs during dialysis within a predetermined time is predicted by the predictive model (S320).

More specifically, the total dialysis time, arterial line pressure (AP), and venous line pressure (VP) obtained from the hemodialysis machine for a predetermined time in the step of receiving the hemodialysis machine data and vital sign data (S310) At least one of a blood flow rate, a dialysate flow rate, an ultrafiltration rate, a total ultrafiltration volume, a dialysis temperature, and a dialysate sodium level may be received.

Furthermore, at least one of systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and pulse rate in the step of receiving hemodialysis machine data and vital sign data (S310) can be received.

At this time, the hemodialysis machine data is a plurality of minute-by-minute data acquired from the hemodialysis machine in units of minutes (eg, hemodialysis machine data acquired by 1 minute for 30 minutes), and the vital sign data is time-units from the vital sign measuring device It may be a plurality of time-unit data (eg, vital sign data acquired in 1-hour units) acquired as .

According to the features of the present invention, in the step of receiving hemodialysis machine data and vital sign data (S310), hemodialysis machine data of blood flow velocity, dialysate sodium level and arterial pressure, and vital sign data of systolic blood pressure and mean arterial pressure may be received. .

Next, in the step of predicting whether hypotension will occur during dialysis (S320), whether hypotension will occur during dialysis within a predetermined time is determined based on the received hemodialysis machine data and vital sign data using the prediction model (S320).

According to a feature of the present invention, in the step of predicting whether hypotension occurs during dialysis (S320), whether hypotension occurs during dialysis within 1 minute to 1 hour can be predicted by a predictive model.

According to another feature of the present invention, in the step of predicting whether hypotension occurs during dialysis (S320), at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' is predicted by the predictive model It can be.

For example, referring to FIG. 4 together, hemodialysis machine data 512 acquired in units of minutes for a predetermined time in the form of segments in the step of predicting whether hypotension occurs during dialysis (S320) and units of time The acquired vital sign data 514 is input into a predictive model 520 . At this time, the predictive model learns to predict the onset of hypotension during dialysis in at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' within a predetermined time by inputting hemodialysis machine data and vital sign data for training. may be a model. As a result, the onset of hypotension 530 during dialysis within a predetermined time period can be predicted and provided. In particular, hypotension during three types of dialysis, 'Nadir90', 'Fall20', and 'Fall20 or MAP10' within a predetermined time (eg, 10 minutes after hemodialysis) can be quickly predicted and provided.

Therefore, according to the information providing method according to various embodiments of the present invention, the medical staff can be provided with information related to hypotension during dialysis of the subject, so that they can make decisions and establish treatment plans more quickly.

That is, by providing an information providing system based on a predictive model learned to predict the occurrence of hypotension during various types of dialysis, medical staff can easily select individualized treatment for each individual.

Therefore, it is possible to predict low blood pressure during dialysis regardless of the skill level of medical staff, and it is possible to provide highly reliable information about it.

Hereinafter, with reference to FIG. 5, structural characteristics of an information prediction model associated with hypotension during dialysis used in various embodiments of the present invention will be described.

FIG. 5 illustratively illustrates the structure of a predictive model used in a method for providing information on hypotension during dialysis according to an embodiment of the present invention.

More specifically, the predictive model 520 used in various embodiments of the present invention includes a convolutional neural network (CNN) module 522 for input of hemodialysis machine data and the vital sign data and an FNN for output of hypotension during dialysis. (feedforward neural network) module 524.

At this time, the CNN module 522 includes an input layer, a computation block 5222 composed of a plurality of computation layers, and an adaptive average pooling (AAP) 5224.

Here, the operation block 5222 may include a one-dimensional convolution (Conv 1D) layer, a one-dimensional average pooling (AvgPooling1D) layer, a batch normalization (BatchNorm) layer, and a scaled exponential linear unit (SELU) activation function layer.

According to an aspect of the present invention, CNN module 522 can consist of one adaptive average pooling (AAP) layer 5224 with four computational blocks 5222.

The FNN module 524 can then include a fully connected layer 5242 and an output layer 5244 .

According to a feature of the present invention, the FNN module 524 comprises two fully connected layers of 128 nodes with three SELU activation functions and one dropout layer with a 0.3 dropout ratio (5242) and an output layer (5244). ) may have a connected structure.

Here, the output layer 5244 may probabilistically predict and output three types of hypotension ('Nadir90', 'Fall20', and 'Fall20 or MAP10') during dialysis through a sigmoid activation function.

According to the above structural characteristics, the predictive model used may probabilistically predict whether hypotension occurs during dialysis within a predetermined time for an individual based on hemodialysis machine data and vital sign data.

On the other hand, the structural characteristics of the predictive model are not limited to the above.

Evaluation 1: Learning and evaluation of predictive models used in various embodiments of the present invention

Hereinafter, results of learning and verification of a predictive model used in various embodiments of the present invention will be described with reference to FIGS. 6A to 6F.

First, referring to FIG. 6A, 26746 hemodialysis subject data from Yonsei University Severance Hospital and 36894 hemodialysis subject data from Myongji University Hospital were used for learning and verification of the predictive model used in various embodiments of the present invention. .

Here, the individual data are total dialysis time, arterial line pressure (AP), venous line pressure (VP), blood flow rate, obtained from the hemodialysis machine in units of 1 minute for 30 minutes flow rate, dialysate flow rate, total ultrafiltration volume, ultrafiltration rate, average ultrafiltration rate, dialysate temperature and dialysate sodium level sodium level) of the hemodialysis machine.

Furthermore, the individual data were acquired from the vital sign measurement device on an hourly basis, systolic blood pressure (SBP) before dialysis, diastolic blood pressure (DBP) before dialysis, mean arterial pressure before dialysis (mean arterial pressure; MAP) and vital sign data of pulse rate.

Here, a total number of segments corresponding to the number of hemodialysis machine data measured in units of 1 minute for 30 minutes may be further included.

More specifically, with reference to FIG. 6B , when data acquisition was performed for 4 hours, the hemodialysis machine data were acquired every minute for 30 minutes between 8:00 and 9:00, arterial line pressure (AP), Includes venous line pressure (VP), dialysate temperature, dialysate sodium level, blood flow rate, dialysate flow rate, and ultrafiltration rate It may be composed of a plurality of segments, such as segment 1 (segment 1), segment 2 including hemodialysis machine data acquired in units of 1 minute for 30 minutes between 9:00 and 10:00.

Furthermore, referring to FIG. 6C , vital sign data such as systolic blood pressure and mean arterial pressure may be acquired every hour for 4 hours.

Next, referring to FIG. 6D, 101655 segments based on Severance Hospital consist of 3.7% of Nadir90, 34.6% of Fall20, and 39% of Fall20 or MAP10 hypotension during dialysis, and 243059 segments based on Myongji University Hospital. The segment consisted of 2.2% Nadir90, 41.1% Fall20, and 45.3% Fall20 or MAP10 hypotension on dialysis.

Referring to FIG. 6E, a logistic regression (LR) based model, a random forest (RF) based model, an extreme gradient boosting (XGB) based model, and a deep learning model ;

More specifically, according to Yonsei University Severance Hospital data-based internal validation, all four algorithm-based predictive models had an average AUROC of 0.89 in predicting hypotension during Nadir90 dialysis. According to Myongji University Hospital data-based external validation, Nadir90's prediction of hypotension during dialysis was shown to have an average AUROC of 0.83 for all four algorithm-based prediction models.

That is, these results may mean that the predictive model trained on the basis of hemodialysis machine data and vital sign data has excellent predictive performance of hypotension during dialysis of Nadir90.

Next, according to internal variables based on Yonsei University Severance Hospital data, in predicting hypotension during dialysis in Fall 20, all four algorithm-based prediction models appear to have an average AUROC of 0.86. According to the external variables based on Myongji University Hospital data, in predicting hypotension during dialysis in Fall 20, all four algorithm-based predictive models appeared to have an average AUROC of 0.86.

That is, these results may mean that the predictive model trained on the basis of hemodialysis machine data and vital sign data has excellent predictive performance of fall 20 hypotension during dialysis.

Next, according to internal variables based on Yonsei University Severance Hospital data, all four algorithm-based predictive models have an average AUROC of 0.85 in predicting hypotension during Fall 20 or MAP10 during dialysis. According to the external variables based on Myongji University Hospital data, in predicting hypotension during Fall 20 or MAP10 during dialysis, all four algorithm-based prediction models appeared to have an average AUROC of 0.84.

That is, these results may mean that the predictive model learned based on the hemodialysis machine data and the vital sign data has excellent performance in predicting hypotension during dialysis in Fall20 or MAP10.

Referring to FIG. 6F, SBP, DPB, MAP and pulse rate-based prediction model (vital signs), SBP, DPB, MAP and pulse rate and AP, VP-based prediction model (vital signs + monitored pressure), SBP, DPB, MAP and pulse rate and blood flow velocity, dialysate flow rate, ultrafiltration rate, total ultrafiltration volume, dialysate temperature and dialysate sodium level-based predictive model (Vital signs + Setting measures), furthermore, SBP, DPB, MAP and derived variables for time information with pulse rate The evaluation results for the predictive model (vital signs + time setting) based on (elapsed time from the start of the hemodialysis session to the start of each segment, elapsed time from the previous SBP measurement to the start of each segment) are shown.

More specifically, in predicting hypotension 10 minutes after 3 types of dialysis, the predictive performance of the predictive model based on vital sign data and hemodialysis machine data is superior to that of the model using only vital sign data.

In particular, in predicting the onset of Nadir90, when the hemodialysis machine data was applied to learning, the prediction performance improved remarkably.

These results indicate that the predictive model based on vital sign data and hemodialysis machine data has superior diagnostic performance in predicting hypotension during dialysis of three types of Nadir90, Fall20, and Fall20 or MAP10 than a single data-based predictive model. there is.

That is, as the information provision system based on the predictive model described above can predict hypotension after a predetermined time (eg, 10 minutes later) during dialysis, medical personnel can quickly perform various treatments.

More specifically, since hypotension can be rapidly predicted during hemodialysis by the predictive model, treatment by medical staff for hypotension, such as the Trendelenburg posture, isotonic solution administration, and ultrafiltration rate reduction, can be effectively performed.

Assessment 2: Assessment of importance of hemodialysis machine data and vital sign data

Hereinafter, results of evaluating the importance of hemodialysis machine data and vital sign data in predicting hypotension during dialysis will be described with reference to FIGS. 7A to 7C .

Referring to FIG. 7A, in predicting Nadir90, vital sign data of systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean arterial pressure (MAP) data are main features. appear.

Referring further to FIG. 7B , in the prediction of Fall20, vital sign data of systolic blood pressure (SBP), mean arterial pressure (MAP) data, hemodialysis machine data of blood flow rate appears as a key feature.

Further referring to FIG. 7C , in the prediction of Fall 20 or MAP 10, systolic blood pressure (SBP), mean arterial pressure (MAP) data, vital sign data, blood flow rate (blood flow rate) Hemodialysis machine data are presented as key features.

That is, in predicting hypotension during the three types of dialysis, vital sign data of systolic blood pressure and mean arterial pressure appear as the most important characteristics. Further, hemodialysis machine data such as blood flow, dialysate sodium level during dialysis and arterial pressure are then featured as key features.

Accordingly, the above-described main data may be used to learn a predictive model for predicting hypotension during dialysis, but is not limited thereto.

Evaluation 3: Assessment of association of hemodialysis machine data and vital signs data with risk of hypotension during dialysis

Hereinafter, a result of evaluating the association of hemodialysis machine data and vital sign data with the risk of hypotension during dialysis will be described with reference to FIGS. 8A to 8C .

At this time, FIGS. 8A to 8C are data showing values measured for 30 minutes as dot plots, and red shading can be interpreted as a direction of increasing the risk of hypotension during dialysis, and blue shading can be interpreted as a direction of lowering the risk.

First, referring to FIG. 8A , it is shown that the change in arterial pressure (AP) increases the risk of Nadir90 occurrence, and the blood flow rate is maintained in the direction of lowering the risk of Nadir90 occurrence. Furthermore, a decrease in dialysate flow rate appears to increase the risk of developing Nadir90. These results may mean that treatment that changes the set values for risk-enhancing variables, namely arterial pressure and dialysis fluid flow rate, can reduce the risk of developing Nadir90.

Next, referring to FIG. 8B , it appears that the fluctuation of arterial pressure (AP) increases the risk of Fall 20, and the blood flow rate appears to be maintained in the direction of lowering the risk of Fall 20. These results may mean that a treatment that changes a variable that increases risk, that is, a set value for arterial pressure, can reduce the risk of Fall20.

Next, referring to FIG. 8C , blood flow rate and dialysate sodium level appear to increase the risk of Fall20 or MAP 10. These results may mean that the treatment of changing the setting values for variables that increase the risk can reduce the risk of Fall20 or MAP10 occurrence.

Although the embodiments of the present invention have been described in more detail with reference to the accompanying drawings, the present invention is not necessarily limited to these embodiments, and may be variously modified and implemented without departing from the technical spirit of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by these embodiments. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the present invention.

100: medical staff device
110: memory interface 120: processor
130 Peripheral Interface 140 I/O Subsystem
141: touch screen controller 142: other input controller
143: touch screen
144: other input control devices
150: memory 151: operating system
152: communication module 153: GUI module
154: sensor processing module 155: phone module
156: applications
156-1, 156-2: application
157: digital assistant client module
158: user data
160: motion sensor 161: light sensor
162 Proximity sensor 163 Other sensors
170: camera subsystem 171: optical sensor
180: communication subsystem
190: audio subsystem
191: speaker 192: microphone
300: server for providing information
310: communication interface
311 wired communication port 312 wireless circuit
320: memory
321: operating system 322: communication module
323: user interface module 324: application
330: I/O interface 340: processor

Claims (22)

A method for providing information on hypotension during hemodialysis implemented by a processor,
receiving hemodialysis machine data and vital sign data obtainable during hemodialysis from the subject; and
Low blood pressure during dialysis based on the received hemodialysis machine data and vital sign data using a prediction model learned to predict hypotension during dialysis within a predetermined time by taking the hemodialysis machine data and the vital sign data as inputs Including the step of predicting,
The hemodialysis machine data is time-series data obtained from the hemodialysis machine in units of minutes, a method for providing information on hypotension during hemodialysis. According to claim 1,
The hemodialysis machine data,
obtained from the hemodialysis machine in minutes for a predetermined time,
Total dialysis time, arterial line pressure (AP), venous line pressure (VP), blood flow rate, dialysate flow rate, ultrafiltration rate, total ultrafiltration volume ( A method for providing information on hypotension during hemodialysis, which is at least one of total ultrafiltration volume), dialysate temperature, and dialysate sodium level. According to claim 1,
The hemodialysis machine data,
Methods for providing information on blood flow rate, dialysate sodium level and arterial pressure, and hypotension during hemodialysis. According to claim 1,
The vital sign data,
A method of providing information on hypotension during hemodialysis, which is at least one of systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and pulse rate. According to claim 1,
The method of providing information on hypotension during hemodialysis, wherein the vital sign data are systolic blood pressure and mean arterial pressure. According to claim 1,
The vital sign data,
A method for providing information on hypotension during hemodialysis, which is time series data obtained in units of time from the subject. According to claim 1,
The predetermined time is from 1 minute to 1 hour, a method for providing information on hypotension during hemodialysis. According to claim 1,
The predicting step is
Using the prediction model, predicting hypotension during dialysis of at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' based on the received hemodialysis machine data and the vital sign data; ,
The Nadir90 is defined as a condition in which the Nadir systolic blood pressure is less than 90 mmHg,
The Fall 20 is defined as a state in which the systolic blood pressure is 20 mmHg or more compared to the initial blood pressure,
The Fall20 or MAP10 is defined as a state in which systolic blood pressure decreases by 20 mmHg or more and/or mean arterial pressure decreases by 10 mmHg or more. According to claim 1,
The predictive model,
A method for providing information on hypotension during hemodialysis, comprising a convolutional neural network (CNN) module for input of the hemodialysis machine data and the vital sign data, and a feedforward neural network (FNN) module for output of hypotension during dialysis. According to claim 9,
The CNN module includes an input layer, a computation block composed of a plurality of computation layers, and an adaptive average pooling layer,
The method of providing information on hypotension during hemodialysis, wherein the FNN module includes an activation function layer, a dropout layer, and an output layer. According to claim 10,
The plurality of computation layers,
A method for providing information on hypotension during hemodialysis, comprising at least two layers of a one-dimensional convolution layer, a one-dimensional average pooling layer, a batch normalization layer, and an activation function layer. A communication unit configured to receive hemodialysis machine data and vital sign data obtainable during hemodialysis from the subject, and
Including a processor functionally connected to the communication unit,
the processor,
Low blood pressure during dialysis based on the received hemodialysis machine data and vital sign data using a prediction model learned to predict hypotension during dialysis within a predetermined time by taking the hemodialysis machine data and the vital sign data as inputs configured to predict
The hemodialysis machine data is time-series data obtained in units of minutes from the hemodialysis machine, a device for providing information on low blood pressure during hemodialysis. According to claim 12,
The hemodialysis machine data,
obtained from the hemodialysis machine in minutes for a predetermined time,
Total dialysis time, arterial line pressure (AP), venous line pressure (VP), blood flow rate, dialysate flow rate, ultrafiltration rate, total ultrafiltration volume ( A device for providing information on hypotension during hemodialysis, which is at least one of a total ultrafiltration volume), a dialysate temperature, and a dialysate sodium level. According to claim 12,
The hemodialysis machine data,
A device for providing information on blood flow rate, dialysate sodium level and arterial pressure, and hypotension during hemodialysis. According to claim 12,
The vital sign data,
A device for providing information on hypotension during hemodialysis, which is at least one of systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and pulse rate. According to claim 12,
The device for providing information on hypotension during hemodialysis, wherein the vital sign data are systolic blood pressure and mean arterial pressure. According to claim 12,
The vital sign data,
A device for providing information on hypotension during hemodialysis, which is time-series data obtained in units of time from the subject. According to claim 12,
The predetermined time is from 1 minute to 1 hour, a device for providing information about hypotension during hemodialysis. According to claim 12,
the processor,
Further configured to predict hypotension during dialysis of at least one of 'Nadir90', 'Fall20', and 'Fall20 or MAP10' based on the received hemodialysis machine data and the vital sign data using the predictive model,
The Nadir90 is defined as a condition in which the Nadir systolic blood pressure is less than 90 mmHg,
The Fall 20 is defined as a state in which the systolic blood pressure is 20 mmHg or more compared to the initial blood pressure,
The Fall20 or MAP10 is a device for providing information on hypotension during hemodialysis, which is defined as a state in which systolic blood pressure decreases by 20 mmHg or more and/or mean arterial pressure decreases by 10 mmHg or more. According to claim 12,
The predictive model,
A device for providing information on hypotension during hemodialysis, comprising a convolutional neural network (CNN) module for input of the hemodialysis machine data and the vital sign data, and a feedforward neural network (FNN) module for output of hypotension during dialysis. According to claim 20,
The CNN module includes an input layer, a computation block composed of a plurality of computation layers, and an adaptive average pooling layer,
The device for providing information on low blood pressure during hemodialysis, wherein the FNN module includes an activation function layer, a dropout layer, and an output layer. According to claim 21,
The plurality of computation layers,
A device for providing information on hypotension during hemodialysis, comprising at least two layers of a one-dimensional convolution layer, a one-dimensional average pooling layer, a batch normalization layer, and an activation function layer.

Images