WO2025053069A1 - Computer program, information processing method, and information processing device - Google Patents

Abstract

Provided are a computer program, an information processing method, and an information processing device.　The present invention acquires information on a peritoneal dialysis patient including at least one of edema, urine output, and ultrafiltration volume, inputs the acquired information to a trained model, which is trained to output information related to a dialysis prescription for the peritoneal dialysis patient when the information on the peritoneal dialysis patient is input, executes computation with the trained model, generates a dialysis prescription on the basis of information output from the trained model, and outputs the generated dialysis prescription.

Description

COMPUTER PROGRAM, INFORMATION PROCESSING METHOD, AND INFORMATION PROCESSING APPARATUS

This disclosure relates to a computer program, an information processing method, and an information processing device.

Various technologies have been proposed to detect the condition of a patient undergoing dialysis treatment. For example, patent literature discloses a technology that divides the period of dialysis treatment into multiple periods and measures the patient's blood pressure during each period.

JP 2000-000217 A

Patent Document 1 makes it possible to manage changes in blood pressure during dialysis, but it does not allow for the proposal of dialysis prescriptions for peritoneal dialysis patients.

In one aspect, the object is to provide a computer program, an information processing method, and an information processing device that can propose dialysis prescriptions for peritoneal dialysis patients.

(1) The computer program disclosed herein is a computer program for causing a computer to execute a process of acquiring information about a peritoneal dialysis patient, including at least one of edema, urine volume, and amount of water removed, inputting the acquired information into a learning model that has been trained to output information related to a dialysis prescription for the peritoneal dialysis patient when the information about the peritoneal dialysis patient is input, executing a calculation using the learning model, generating a dialysis prescription based on the information output from the learning model, and outputting the generated dialysis prescription.

(2) In the computer program of (1) above, it is preferable that the dialysis prescription includes at least one of the type of peritoneal dialysis fluid to be prescribed and the number of peritoneal dialysis sessions per day.

(3) In the computer program of (1) or (2) above, it is preferable that the dialysis prescription includes replacing a currently prescribed peritoneal dialysis fluid with a different type of peritoneal dialysis fluid.

(4) In any one of the computer programs (1) to (3) above, it is preferable to have the computer execute a process of generating an image showing the bodily fluid status of a peritoneal dialysis patient based on the acquired information about the patient, and outputting the generated image.

(5) In the computer program of (4) above, it is preferable to have the computer execute a process for generating, as the image, an image that shows a schematic representation of the amount of sodium stored and the amount of sodium excreted in the peritoneal dialysis patient.

(6) In any one of the computer programs (1) to (5) above, it is preferable that the information further includes at least one of information regarding hANP (human atrial natriuretic peptide), cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion.

(7) In any one of the computer programs (1) to (6) above, it is preferable that the information on the peritoneal dialysis patient further includes information on at least one of the body weight, dry mouth, antihypertensive drugs, diuretics, and blood pressure of the peritoneal dialysis patient.

(8) In any one of the computer programs (1) to (7) above, it is preferable that the information on the peritoneal dialysis patient further includes information on at least one of urinary cortisol, urinary creatinine, skin sodium, and Na-MRI (nuclear magnetic resonance).

(9) In any one of the computer programs (1) to (8) above, it is preferable that the learning model is a language generation model trained to output sentences related to peritoneal dialysis prescriptions, and the computer is caused to execute a process of outputting the sentences related to the peritoneal dialysis prescriptions by inputting acquired information into the language generation model.

(10) The information processing method disclosed herein acquires information about a peritoneal dialysis patient, including at least one of edema, urine volume, and amount of water removed, inputs the acquired information into a learning model that has been trained to output information related to a dialysis prescription for the peritoneal dialysis patient when the information about the peritoneal dialysis patient is input, executes a calculation using the learning model, generates a dialysis prescription based on the information output from the learning model, and executes a process of outputting the generated dialysis prescription by a computer.

(11) The information processing device disclosed herein includes at least one processor, which acquires information about a peritoneal dialysis patient including at least one of edema, urine volume, and amount of water removed, inputs the acquired information into a learning model that has been trained to output information related to a dialysis prescription for the peritoneal dialysis patient when the information about the peritoneal dialysis patient is input, executes a calculation using the learning model, generates a dialysis prescription based on the information output from the learning model, and outputs the generated dialysis prescription.

On one hand, it can suggest dialysis prescriptions for peritoneal dialysis patients.

1 is a schematic diagram showing a configuration example of a peritoneal management support system according to an embodiment; FIG. 2 is a block diagram illustrating an internal configuration of the server device. FIG. 2 is a block diagram illustrating an internal configuration of a patient terminal. FIG. 2 is a block diagram illustrating the internal configuration of a doctor terminal. FIG. 13 is a schematic diagram showing an example of a patient information input screen on a patient terminal. 13 is a schematic diagram showing an example of a patient information input screen on a doctor terminal. FIG. FIG. 2 is a schematic diagram showing an example of the configuration of a learning model. FIG. 2 is a schematic diagram showing an example of a dialysis prescription generated by the server device. 11 is a flowchart illustrating a procedure of a process executed by a server device. FIG. 2 is a schematic diagram showing an example of the configuration of a learning model. FIG. 13 is a schematic diagram showing an example of image generation. FIG. 13 is a schematic diagram showing an example of image generation. FIG. 13 is a schematic diagram showing an example of image generation. FIG. 11 is a schematic diagram showing an example of output in the second embodiment. FIG. 11 is a schematic diagram showing an example of output in the second embodiment.

The present invention will now be described in detail with reference to the drawings showing embodiments thereof.
(Embodiment 1)
1 is a schematic diagram showing a configuration example of a peritoneal management support system according to an embodiment. The peritoneal management support system according to the embodiment includes a server device 10, a patient terminal 20, and a doctor terminal 30. The server device 10 is a server computer installed, for example, by a hospital. The patient terminal 20 is a client computer used by a peritoneal dialysis patient (hereinafter also simply referred to as a patient), and the doctor terminal 30 is a client computer used by a medical professional such as a doctor. The server device 10, the patient terminal 20, and the doctor terminal 30 are connected to each other so as to be able to communicate with each other via a communication network NW.

The server device 10 acquires information on peritoneal dialysis patients (hereinafter referred to as patient information) from the patient terminal 20 and the doctor terminal 30. The patient information acquired by the server device 10 includes information on at least one of edema, urine volume, and amount of water removed. Information on edema, urine volume, and amount of water removed can be received through the patient terminal 20. The patient information acquired by the server device 10 may further include information such as dry mouth and weight input from the patient terminal 20.

The patient information acquired by the server device 10 may further include information regarding at least one of hANP, cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion. This information is managed for each patient, and may be entered, for example, from the doctor's terminal 30 and preregistered in the server device 10. The patient information registered in the server device 10 may further include information regarding antihypertensive drugs, diuretics, blood pressure, etc., and may further include information regarding urinary cortisol, urinary creatinine, skin sodium, Na-MRI, etc.

The server device 10 generates a dialysis prescription for a peritoneal dialysis patient in response to input of patient information, and outputs the generated dialysis prescription. For example, the server device 10 displays the generated dialysis prescription on the display screen of the server device 10. Alternatively, the server device 10 may transmit the generated dialysis prescription to the patient terminal 20 or to the doctor terminal 30 via the communication network NW.

The internal structure of each device that makes up the peritoneal management support system is explained below.

FIG. 2 is a block diagram explaining the internal configuration of the server device 10. The server device 10 is a dedicated or general-purpose computer, and includes a control unit 11, a storage unit 12, a communication unit 13, an operation unit 14, and a display unit 15.

The control unit 11 includes, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The ROM included in the control unit 11 stores control programs and the like that control the operation of each piece of hardware included in the server device 10. The CPU in the control unit 11 reads and executes the control programs stored in the ROM and computer programs (described below) stored in the memory unit 12, and controls the operation of each piece of hardware, causing the entire device to function as the server device 10 (information processing device) of the present disclosure. The RAM included in the control unit 11 temporarily stores data used during the execution of calculations and control.

In the embodiment, the control unit 11 is configured to include a CPU, ROM, and RAM, but the configuration of the control unit 11 is not limited to the above. The control unit 11 may be, for example, one or more control circuits or arithmetic circuits including a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), a quantum processor, volatile or non-volatile memory, etc. The control unit 11 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The memory unit 12 includes a storage device such as a hard disk drive (HDD) or a solid state drive (SSD). The memory unit 12 stores various computer programs executed by the control unit 11 and various data used by the control unit 11.

The computer program (program product) stored in the memory unit 12 includes a prescription generation program PG1 for causing a computer to execute a process of generating a dialysis prescription to be prescribed to a patient using a learning model LM1 described below and outputting the generated dialysis prescription. The prescription generation program PG1 may be a single computer program or a group of programs consisting of multiple computer programs. The prescription generation program PG1 may also partially use an existing library. The prescription generation program PG1 may be executed by a single computer or may be executed by multiple computers working together.

The computer program including the prescription generation program PG1 is provided by a non-transitory recording medium RM on which the computer program is recorded in a readable manner. The recording medium RM is a portable memory such as a CD-ROM, a USB memory, or an SD (Secure Digital) card. The control unit 11 reads various computer programs from the recording medium RM using a reading device not shown in the figure, and stores the various computer programs that have been read in the memory unit 12. The computer program including the prescription generation program PG1 may be provided by communication. The control unit 11 acquires the computer program including the prescription generation program PG1 by communication via the communication unit 13, and stores the acquired computer program in the memory unit 12.

The memory unit 12 also stores a learning model LM1 that has been trained to output information related to a dialysis prescription for a peritoneal dialysis patient (hereinafter also referred to as prescription information) when information related to at least one of edema, urine volume, and water removal volume is input. Specifically, the memory unit 12 stores information describing the learned learning model LM1, such as the layer structure of the model, the number and connection relationships of nodes that make up each layer, and weighting coefficients and biases between nodes, as well as trained parameters, for the trained learning model LM1.

The communication unit 13 has a communication interface for transmitting and receiving various data to and from an external device. As the communication interface of the communication unit 13, a communication interface conforming to a communication standard such as WiFi (registered trademark), LAN (Local Area Network), Bluetooth (registered trademark), ZigBee (registered trademark), 3G, 4G, 5G, or LTE (Long Term Evolution) can be used. When data to be transmitted is input from the control unit 11, the communication unit 13 transmits the data to the destination external device, and when data transmitted from an external device is received, the communication unit 13 outputs the received data to the control unit 11. In this embodiment, examples of external devices are the patient terminal 20 and the doctor terminal 30.

The operation unit 14 is equipped with operation devices such as a touch panel, a keyboard, and switches, and accepts various operations and settings by the user, etc. The control unit 11 performs appropriate control based on the various operation information provided by the operation unit 14, and stores the setting information in the storage unit 12 as necessary.

The display unit 15 includes a display device such as a liquid crystal display or an organic EL (Electro-Luminescence) display, and displays the information to be presented in response to instructions from the control unit 11.

In this embodiment, the server device 10 may be a single computer, or may be a computer system composed of multiple computers and peripheral devices. In addition, the server device 10 may be a virtual machine whose entity has been virtualized, or may be a cloud.

FIG. 3 is a block diagram explaining the internal configuration of the patient terminal 20. The patient terminal 20 is a computer such as a smartphone, a tablet terminal, or a personal computer, and includes a control unit 21, a memory unit 22, a communication unit 23, an operation unit 24, and a display unit 25.

The control unit 21 includes, for example, a CPU, a ROM, a RAM, etc. The CPU of the control unit 21 controls the operation of the various hardware components described above by expanding into the RAM various programs pre-stored in the ROM or the storage unit 22 and executing them, causing the entire device to function as the patient terminal 20 of the present disclosure.

The control unit 21 is not limited to the above configuration, and may be configured as a single piece of hardware (SoC: System On a Chip) that integrates a processor, memory, storage, communication interface, etc. The control unit 21 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The storage unit 22 includes storage such as a memory and a hard disk. The storage unit 22 stores various computer programs executed by the control unit 21, and data necessary for executing the computer programs. The computer programs stored in the storage unit 22 include application programs for using the services provided by the server device 10.

The programs stored in the memory unit 22 may be provided by a non-transitory recording medium on which the programs are recorded in a readable manner. The recording medium is, for example, a portable memory such as a CD-ROM, a USB memory, or an SD card. The control unit 21 reads the various programs from the recording medium using a reading device (not shown), and installs the various programs that have been read into the memory unit 22. The programs stored in the memory unit 22 may also be provided by communication. In this case, the control unit 21 acquires the various programs through the communication unit 23, and installs the acquired programs into the memory unit 22.

The communication unit 23 has a communication interface for connecting to the communication network NW. The communication interface of the communication unit 23 is, for example, a communication interface conforming to a communication standard such as WiFi (registered trademark), LAN, Bluetooth (registered trademark), ZigBee (registered trademark), 3G, 4G, 5G, or LTE. The communication unit 23 transmits various information to be notified to the outside, and receives various information transmitted from the outside to the device itself.

The operation unit 24 is equipped with input devices such as a touch panel and operation buttons, and receives various types of operation information and setting information. The operation unit 24 outputs the received operation information and setting information to the control unit 21. The control unit 21 performs appropriate control based on the operation information input from the operation unit 24, and stores the setting information in the memory unit 22 as necessary.

The display unit 25 is equipped with a display device such as a liquid crystal display or an organic EL display, and displays information to be presented to the patient based on the control signal output from the control unit 21.

FIG. 4 is a block diagram explaining the internal configuration of the doctor terminal 30. The doctor terminal 30 is a computer such as a personal computer, a tablet terminal, or a smartphone, and includes a control unit 31, a memory unit 32, a communication unit 33, an operation unit 34, and a display unit 35.

The control unit 31 includes, for example, a CPU, a ROM, a RAM, etc. The CPU of the control unit 31 controls the operation of the various hardware components described above by expanding into the RAM various programs pre-stored in the ROM or the storage unit 32 and executing them, causing the entire device to function as the doctor terminal 30 of the present disclosure.

The control unit 31 is not limited to the above configuration, and may be configured as a single piece of hardware (SoC: System On a Chip) that integrates a processor, memory, storage, communication interface, etc. The control unit 31 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The storage unit 32 includes storage such as a memory and a hard disk. The storage unit 32 stores various computer programs executed by the control unit 31, and data necessary for executing the computer programs. The computer programs stored in the storage unit 32 include application programs for using the services provided by the server device 10.

The programs stored in the memory unit 32 may be provided by a non-transitory recording medium on which the programs are recorded in a readable manner. The recording medium is, for example, a portable memory such as a CD-ROM, a USB memory, or an SD card. The control unit 31 reads the various programs from the recording medium using a reading device (not shown), and installs the various programs that have been read into the memory unit 32. The programs stored in the memory unit 32 may also be provided by communication. In this case, the control unit 31 acquires the various programs through the communication unit 33, and installs the acquired programs into the memory unit 32.

The communication unit 33 has a communication interface for connecting to the communication network NW. The communication interface of the communication unit 33 is, for example, a communication interface conforming to communication standards such as WiFi (registered trademark), LAN, Bluetooth (registered trademark), ZigBee (registered trademark), 3G, 4G, 5G, and LTE. The communication unit 33 transmits various information to be notified to the outside, and receives various information transmitted from the outside to its own device.

The operation unit 34 is equipped with input devices such as a keyboard and a mouse, and receives various types of operation information and setting information. The operation unit 34 outputs the received operation information and setting information to the control unit 31. The control unit 31 performs appropriate control based on the operation information input from the operation unit 34, and stores the setting information in the memory unit 32 as necessary.

The display unit 35 is equipped with a display device such as a liquid crystal display or an organic EL display, and displays information to be presented to medical personnel based on the control signal output from the control unit 31.

The operation of the peritoneal management support system will now be described.
When creating a dialysis prescription, the peritoneal management support system according to the present embodiment acquires patient information from the patient terminal 20 used by the patient. The patient information is input by the patient himself/herself through a patient information input screen shown below.

FIG. 5 is a schematic diagram showing an example of a patient information input screen on the patient terminal 20. The patient information input screen shown in FIG. 5 shows an example of a screen that is displayed on the display unit 25 when the server device 10 is accessed from the patient terminal 20 and a specified operation is performed. This patient information input screen accepts information regarding edema, urine volume, and amount of water removed.

The state of edema is measured by the patient checking the indentation. Specifically, the area where bones such as the tibia are subcutaneously placed is pressed with the thumb or the second to fourth fingers, and the condition of the skin after the pressure is released is checked visually and palpably. An edema level of "0" indicates a normal state with no indentation. An edema level of "1+" indicates a slight indentation, but no visible distortion, and the indentation disappears quickly. An edema level of "2+" indicates a somewhat deep indentation, but the indentation disappears within about 10 to 15 seconds. An edema level of "3+" indicates a deep indentation that may last for one minute or more. An edema level of "4+" indicates a very deep indentation that lasts for two to five minutes or more. In the patient information input screen in Figure 5, the state of edema is entered into three categories: "0 to 1+", "2+", and "3+ to 4+". Alternatively, the edema condition may be entered into five categories ranging from "0" to "4+".

Urine volume is measured by the patient themselves using a measuring cup or similar. The daily urine volume (mL) is entered on the patient information input screen in Figure 5. The amount of water removed is calculated as the amount of fluid injected - the amount of fluid drained. Here, the amount of fluid injected represents the amount of dialysis fluid injected into the patient's body, and is measured by weighing the bag containing the dialysis fluid before and after injection. The amount of fluid drained represents the amount of dialysis fluid discharged from the patient's body, and is measured by weighing the bag that collects the drained fluid. The daily amount of water removed (mL) is entered on the patient information input screen in Figure 5.

In FIG. 5, an example of a patient information input screen that accepts information regarding edema, urine volume, and amount of water removed is shown, but it may also be possible to accept any one or any two of edema, urine volume, and amount of water removed.

Furthermore, information on dry mouth or body weight may be received on the patient information input screen. Dry mouth is subjectively rated by the patient on a numerical rating scale (NRS) of 0 to 10. Alternatively, dry mouth may be rated on a two-level scale: dry mouth/no dry mouth. Body weight is measured by the patient using a weighing scale.

Furthermore, some of the patient information may be entered by a medical professional. Figure 6 is a schematic diagram showing an example of a patient information input screen on the doctor terminal 30. The patient information input screen shown in Figure 6 shows an example of a screen that is displayed on the display unit 35 when the server device 10 is accessed from the doctor terminal 30 and a specified operation is performed. This patient information input screen accepts information related to hANP, cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion.

hANP is a human atrial natriuretic peptide, and its concentration is measured by a blood test. The IRMA (immunoradiometric assay) method is used to measure the concentration. The reference value for hANP is, for example, 43.0 pg/mL. hAMP is secreted in response to stimulation of the atrium, so it shows abnormally high values in diseases or pathological conditions that cause increased atrial pressure or increased body fluid volume. In the example screen in Figure 6, the hAMP concentration is entered as a number (pg/mL).

The cardiothoracic ratio is an index that shows the ratio of the transverse diameter of the heart to the transverse diameter of the thorax, and is measured using a chest X-ray. The appropriate ratio is 50% or less for men and 55% or less for women. When the amount of extracellular fluid increases, the amount of water in the blood vessels increases and the heart becomes larger, so the cardiothoracic ratio increases. In the example screen in Figure 6, the cardiothoracic ratio is entered as a number (%).

Bioimpedance is the resistance (impedance) of a living body, and is measured by a body composition analyzer. Body composition analyzers use the bioelectrical impedance analysis (BIA) method, which takes advantage of the fact that almost no current flows through fat tissue, but that current flows easily through lean tissue, which contains water and electrolytes. By using this BIA method, body composition analyzers can measure the indicator of excess/deficiency of body fluids (OH). The standard value for OH is -1.1 to 1.1 L. In the example screen in Figure 6, the indicator of excess/deficiency of body fluids is entered as a numerical value (L).

The transperitoneal sodium removal amount is measured, for example, by collecting the dialysis fluid (effluent) discharged from the patient's body during peritoneal dialysis and analyzing the components of the collected effluent. In the example screen in Figure 6, the transperitoneal sodium removal amount is entered as a value (g) converted to salt.

The amount of urinary sodium removal is measured, for example, by collecting urine from the patient and analyzing the components of the collected urine. In the example screen of Figure 6, the amount of urinary sodium removal is entered as a value (g) converted to salt.

In FIG. 6, an example of a patient information input screen that accepts information related to hANP, cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion is shown, but any one or more of these pieces of information may be accepted. Also, it is not necessary to accept all information on a single input screen, and each piece of information may be accepted individually.

Furthermore, the patient information input screen may accept information regarding at least one of antihypertensive drugs, diuretics, blood pressure, urinary cortisol, urinary creatinine, skin sodium, and Na-MRI.

The patient information input to the patient terminal 20 and the doctor terminal 30 is transmitted to the server device 10 via the communication network NW. The server device 10 generates a dialysis prescription for the patient based on the patient information acquired from the patient terminal 20 and the doctor terminal 30. In this embodiment, the patient information is input to the learning model LM1, and a dialysis prescription is generated based on the prescription information obtained by performing calculations using the learning model LM1.

Figure 7 is a schematic diagram showing an example of the configuration of the learning model LM1. When patient information is input, the learning model LM1 is trained to output prescription information for that patient. Known learning models such as CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), LSTM (Long Short Term Memory), and autoencoder are used for the learning model LM1.

The learning model LM1 includes an input layer, an intermediate layer, and an output layer. Each layer includes one or more nodes. Patient information is input to the nodes of the input layer. Information on at least one of edema, urine volume, and water removal volume is input to the nodes of the input layer. The information input to the nodes of the input layer may further include at least one of hANP, cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion. The information input to the nodes of the input layer may also include information on at least one of the patient's weight, dry mouth, antihypertensive drugs, diuretics, and blood pressure. The information input to the nodes of the input layer may also include information on at least one of urinary cortisol, urinary creatinine, skin sodium, and Na-MRI.

The intermediate layer of the learning model LM1 is composed of one or more layers. Each node in the intermediate layer is connected to the nodes in the layer before and after it. The activity level of each node and the strength of the connection between nodes (weighting coefficient) are determined during the learning process. The nodes in each layer output values calculated based on the weighting coefficient and activity level to the next node.

The output layer calculates probabilities using a softmax function based on values input from the intermediate layer, and outputs information (prescription information) based on the calculated probabilities. For example, the output layer calculates the probability that the type of peritoneal dialysis fluid to be prescribed is "liquid A", "liquid B", ..., "liquid X", and outputs information based on the calculated probabilities. Alternatively, the output layer may calculate the probability that the number of peritoneal dialysis sessions on the Nth day (N is a natural number) is 1 session, 2 sessions, ..., n sessions (n is, for example, 4), and output information based on the calculated probabilities. The output layer may also output both information related to the type of peritoneal dialysis and information related to the number of sessions per day.

Such a learning model LM1 is generated by acquiring a dataset containing patient information and a dialysis prescription (correct answer data) prescribed by a doctor for a patient who has undergone appropriate peritoneal dialysis, and using the acquired dataset as training data to learn with an existing learning algorithm. The memory unit 12 of the server device 10 stores the learned learning model LM1.

When the control unit 11 of the server device 10 acquires patient information from the patient terminal 20 or the doctor terminal 30, it inputs the acquired patient information into the learning model LM1, executes a calculation, and acquires prescription information as the calculation result by the learning model LM1.

The control unit 11 of the server device 10 generates a dialysis prescription for the patient based on the prescription information output from the learning model LM1. Figure 8 is a schematic diagram showing an example of a dialysis prescription generated by the server device 10. Figure 8 shows the progress of the dialysis prescription over the course of peritoneal dialysis therapy. In the example of FIG. 8, it is shown that in period A (period when urine volume is sufficient), peritoneal dialysis using "low glucose concentration peritoneal dialysis solution" is performed twice a day, in period B (period when the amount of water removed is to be increased due to a tendency of excess body fluid), the number of times of peritoneal dialysis using "low glucose concentration peritoneal dialysis solution" is increased to three times, in period C (period when urine volume has decreased slightly), the number of times of peritoneal dialysis using "low glucose concentration peritoneal dialysis solution" is increased to four times, in period D (period when the tendency of the amount of water removed to decrease is accelerating), "low glucose concentration peritoneal dialysis solution" is sequentially replaced with "medium glucose concentration peritoneal dialysis solution", and in period E (period when urine volume is insufficient and the amount of water removed is insufficient with only medium glucose concentration peritoneal dialysis solution), the first "medium glucose concentration peritoneal dialysis solution" is replaced with "icodextrin-containing peritoneal dialysis solution". Also shown is an estimated time point of "present" in a long-term peritoneal dialysis therapy period. In the example of Figure 8, the trends in the amount of urine and the amount of water removed per day are also shown. These values are estimated values when peritoneal dialysis therapy is performed as shown in the figure, and show values calculated by the control unit 11 according to the type of peritoneal dialysis fluid and the amount of fluid injected.

The control unit 11 generates a dialysis prescription as shown in FIG. 8 as a document with an image, and outputs the generated dialysis prescription to the outside. The output destination may be the display unit 15 of the information processing device 10, or the patient terminal 20 or the doctor terminal 30. The dialysis prescription shown in FIG. 8 represents a dialysis prescription that complies with incremental PD (peritoneal dialysis). Incremental PD is a stepwise peritoneal dialysis introduction method in which renal replacement therapy is started with a prescription for fewer PD bag exchanges in the early stages of introduction while residual renal function is maintained, and the amount of dialysis is increased according to the degree of decline in residual renal function.

In FIG. 8, an example of a dialysis prescription including the type of peritoneal dialysis fluid to be prescribed and the number of peritoneal dialysis sessions per day is shown, but the control unit 11 may generate a dialysis prescription including either the type of peritoneal dialysis fluid to be prescribed or the number of peritoneal dialysis sessions per day. In addition, instead of the above dialysis prescription, the control unit 11 may generate a dialysis prescription (see FIG. 12) including replacing the currently prescribed peritoneal dialysis fluid with another type of peritoneal dialysis fluid, and output the generated dialysis prescription.

FIG. 9 is a flowchart explaining the procedure of the process executed by the server device 10. The server device 10 acquires patient information from the patient terminal 20 and the doctor terminal 30 via the communication network NW (step S101). The patient information acquired by the server device 10 includes information on at least one of edema, urinary volume, and water removal volume. In addition to this information, information such as hANP, cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion may also be acquired.

The control unit 11 of the server device 10 inputs the acquired patient information into the learning model LM1 and executes a calculation using the learning model LM1 (step S102). The control unit 11 acquires prescription information as a result of the calculation using the learning model LM1 (step S103).

The control unit 11 generates a dialysis prescription for the patient based on the acquired prescription information (step S104). The calculation results by the learning model LM1 include at least one of information on the type of peritoneal dialysis fluid to be prescribed for the patient and the number of peritoneal dialysis sessions per day, so the control unit 11 can generate a dialysis prescription based on this information. In one example, the control unit 11 can generate a dialysis prescription as shown in FIG. 8.

The control unit 11 outputs the generated dialysis prescription (step S105). For example, the control unit 11 can display the generated dialysis prescription on the display unit 15. The control unit 11 may also transmit the generated dialysis prescription to the patient terminal 20 or the doctor terminal 30. The dialysis prescription transmitted from the server device 10 reaches the patient terminal 20 or the doctor terminal 30 through the communication network NW. When the patient terminal 20 receives the dialysis prescription transmitted from the server device 10, it displays the received dialysis prescription on the display unit 25. When the doctor terminal 30 receives the dialysis prescription transmitted from the server device 10, it displays the received dialysis prescription on the display unit 35.

As described above, in this embodiment, a dialysis prescription for a patient can be presented to the patient himself or herself or to medical staff. By checking the screen displayed on the display unit 25 of the patient terminal 20, the patient can ascertain the appropriate dialysis prescription according to his or her bodily fluid status. By checking the screen displayed on the display unit 35 of the doctor terminal 30, medical staff can use it as a community tool to provide guidance on appropriate peritoneal management.

In this embodiment, the server device 10 is configured to send the dialysis prescription directly to the patient terminal 20, but it may also be configured to send the prescription to the doctor terminal 30 before sending it to the patient terminal 20, and then send it to the patient terminal 20 after receiving approval from the doctor.

In this embodiment, the server device 10 is configured to execute calculations using the learning model LM1 and generate a dialysis prescription using the calculation results of the learning model LM1, but the patient terminal 20 and the doctor terminal 30 may each execute calculations using the learning model LM1 and generate a dialysis prescription using the calculation results of the learning model LM1. In this case, the prescription generation program PG1 and the learning model LM1 are installed in each of the patient terminal 20 and the doctor terminal 30.

Also, a language generation model may be used as the learning model LM1. The learning model LM1 based on the language generation model is trained to output sentences related to a peritoneal dialysis prescription in response to input of patient information. Such a learning model LM1 is generated, for example, by fine-tuning an existing large-scale language model. The learning model LM1 based on the language generation model outputs sentences such as "At this time, please perform dialysis using low-glucose concentration peritoneal dialysis fluid once a day. As the amount of urine is decreasing, please consider performing dialysis using low-glucose concentration peritoneal dialysis fluid twice a day at the next outpatient visit..." in response to input of patient information. The control unit 11 of the server device 10 generates a dialysis prescription based on the sentences output from the learning model LM1, and outputs the dialysis prescription by displaying it on the display unit 15 or transmitting it to the patient terminal 20 or the doctor terminal 30.

(Embodiment 2)
In the second embodiment, a configuration is described in which an image showing the state of the patient's bodily fluids is generated and output together with a dialysis prescription. The overall configuration of the peritoneal management support system and the internal configuration of each device constituting the system are the same as those in the first embodiment, and therefore the description thereof is omitted.

The server device 10 according to the second embodiment generates an image showing the bodily fluid status of a patient based on patient information. For example, when the server device 10 receives patient information, it can generate an image showing the bodily fluid status using a learning model LM2 that has been trained to output information about the patient's bodily fluid status.

FIG. 10 is a schematic diagram showing an example of the configuration of the learning model LM2. When patient information is input, the learning model LM2 is trained to output information on the patient's bodily fluid status. As with the learning model LM1 described in embodiment 1, the learning model LM2 uses known learning models such as CNN, RNN, LSTM, and autoencoder. The learning model LM2 has an input layer, an intermediate layer, and an output layer, and performs calculations in the intermediate layer in response to patient information input to the input layer, and outputs information on the bodily fluid status from the output layer. For example, the output layer calculates the probability that the bodily fluid volume is normal, the probability that it is excessive, and the probability that it is excessive, and outputs information based on the probability. For example, the learning model LM2 may output information that determines whether the patient's bodily fluid volume is normal, tends to be excessive, or is excessive in response to input of patient information. Such a learning model LM2 is generated by learning with an existing learning algorithm using a data set as training data that includes patient information such as edema, urine volume, and amount of water removed obtained from the patient, and actual values (correct values) of bodily fluid volume measured using a measuring device such as a body composition analyzer. The memory unit 12 of the server device 10 stores the learned learning model LM2.

When the control unit 11 of the server device 10 acquires patient information from the patient terminal 20 or the doctor terminal 30, it inputs the acquired patient information into the learning model LM2 to execute a calculation, and acquires the calculation result by the learning model LM2. In the second embodiment, the calculation result by the learning model LM2 is information that determines whether the patient's body fluid volume is normal, tends to be excessive, or is excessive.

The control unit 11 of the server device 10 generates an image showing the patient's bodily fluid status based on the calculation results of the learning model LM2. Figures 11A to 11C are schematic diagrams showing examples of image generation. Each image depicts a first container V1 that contains interstitial fluid, a second container V2 that contains plasma, a flow path P1 that connects the first container V1 and the second container, and an excretory path P2 that discharges liquid from the second container. The interstitial fluid contained in the first container V1 and the plasma contained in the second container V2 can flow from one container to the other through the flow path P1. The amount of sodium excreted as urine is depicted as the amount of liquid discharged from this excretory path P2. The amount of sodium removed by dialysis is depicted by the number of ">" symbols.

FIG. 11A shows that sodium removal by peritoneal dialysis is sufficient, and bodily fluid management is good. If the learning model LM2 determines that the patient's bodily fluid volume is normal, the control unit 11 generates an image as shown in FIG. 11A.

FIG. 11B shows that sodium removal by peritoneal dialysis is insufficient, and bodily fluid management is somewhat poor. If the learning model LM2 determines that the patient's bodily fluid volume is tending to be excessive, the control unit 11 generates an image showing that the interstitial fluid volume and plasma volume are somewhat high, as shown in FIG. 11B.

FIG. 11C shows that sodium removal by peritoneal dialysis is not working properly, and that bodily fluid management is poor. If the learning model LM2 determines that the patient has an excess of bodily fluids, the control unit 11 generates an image of the first container V1 and the second container V2 filled with interstitial fluid and plasma, respectively, as shown in FIG. 11C.

In this embodiment, patient information is input into learning model LM2 to obtain information regarding the bodily fluid status, and an image representing the patient's bodily fluid status is generated based on the obtained information. However, it is also possible to use an image generation learning model such as GAN (Generative Adversarial Network) or PGGAN (Progressive Growing of GANs) to generate an image representing the patient's bodily fluid status directly from patient information.

In the second embodiment, an image showing the generated fluid status of the patient is output together with the dialysis prescription. Figs. 12 and 13 are schematic diagrams showing examples of output in the second embodiment. Fig. 12 shows an example of a dialysis prescription generated for a patient with good fluid management. This example shows a dialysis prescription that includes replacing the peritoneal dialysis fluid in the current prescription (liquid A in the example of Fig. 12) with a different type of peritoneal dialysis fluid (low Na solution in the example of Fig. 12). In addition, in the example of Fig. 12, a sentence is shown as a summary of the fluid management, stating that although there is no fluid excess, the amount of Na accumulation has increased slightly, and that it is proposed to replace liquid A with low Na solution. This sentence may be a standard phrase, or may be generated by a language generation model.

FIG. 13 shows an example of a dialysis prescription generated for a patient with somewhat poor fluid management. This example shows a dialysis prescription that includes replacing the peritoneal dialysis fluid in the current prescription (liquid B in the example of FIG. 13) with a different type of peritoneal dialysis fluid (icodextrin solution in the example of FIG. 13). The example of FIG. 13 also shows a summary of fluid management, stating that there is a tendency for fluid excess, that sodium accumulation is increasing, and that it is suggested to replace liquid B with icodextrin solution. This sentence may be a standard phrase, or may be generated by a language generation model.

As described above, in the second embodiment, a dialysis prescription is presented to the patient along with an image showing the patient's bodily fluid status, so the patient can intuitively understand their own bodily fluid status by checking the image displayed on the display unit 25 of the patient terminal 20. Medical staff can grasp the patient's bodily fluid status by checking the image displayed on the display unit 35 of the doctor terminal 30, and can use it as a community tool to provide guidance on appropriate bodily fluid management.

The embodiments disclosed herein are illustrative in all respects and should not be considered limiting. The scope of the present invention is indicated by the claims, not by the meaning described above, and is intended to include all modifications within the meaning and scope of the claims.

REFERENCE SIGNS LIST 10 Server device 11 Control unit 12 Memory unit 13 Communication unit 14 Operation unit 15 Display unit 20 Patient terminal 21 Control unit 22 Memory unit 23 Communication unit 24 Operation unit 25 Display unit 30 Doctor terminal 31 Control unit 32 Memory unit 33 Communication unit 34 Operation unit 35 Display unit PG1 Prescription generation program LM1 Learning model

Claims (11)

Acquire information on a peritoneal dialysis patient, the information including at least one of edema, urine volume, and water removal volume;
inputting the acquired information into a learning model that has been trained to output information related to a dialysis prescription for a peritoneal dialysis patient when the information of the peritoneal dialysis patient is input, and executing a calculation by the learning model;
generating a dialysis prescription based on the information output from the learning model;
A computer program for causing a computer to execute a process of outputting the generated dialysis prescription. The computer program product of claim 1 , wherein the dialysis prescription includes at least one of a type of peritoneal dialysis fluid to be prescribed and a number of peritoneal dialysis sessions per day. The computer program product of claim 1 , wherein the dialysis prescription includes substituting a different type of peritoneal dialysis fluid for a currently prescribed peritoneal dialysis fluid. generating an image representing a bodily fluid status of the peritoneal dialysis patient based on the acquired information of the peritoneal dialysis patient;
The computer program product according to claim 1 , for causing the computer to execute a process of: outputting the generated image. The computer program product according to claim 4, for causing the computer to execute a process of generating, as the image, an image that diagrammatically represents sodium storage and sodium excretion in the peritoneal dialysis patient. The computer program product according to claim 1 , wherein the information on the peritoneal dialysis patient further includes information on at least one of human atrial natriuretic peptide (hANP), cardiothoracic ratio, bioimpedance, transperitoneal sodium removal, and urinary sodium excretion. The computer program product according to claim 6 , wherein the information on the peritoneal dialysis patient further includes information on at least one of the weight, dry mouth, antihypertensive drug, diuretic drug, and blood pressure of the peritoneal dialysis patient. 8. The computer program according to claim 7, wherein the information on the peritoneal dialysis patient further includes information on at least one of urinary cortisol, urinary creatinine, skin sodium, and Na-MRI (nuclear magnetic resonance). the learning model is a language generation model trained to output sentences related to a peritoneal dialysis prescription,
The computer program product according to claim 1 , which causes the computer to execute a process of: inputting acquired information into the language generation model, thereby outputting a sentence related to the peritoneal dialysis prescription. Acquire information on a peritoneal dialysis patient, the information including at least one of edema, urine volume, and water removal volume;
inputting the acquired information into a learning model that has been trained to output information related to a dialysis prescription for a peritoneal dialysis patient when the information of the peritoneal dialysis patient is input, and executing a calculation by the learning model;
generating a dialysis prescription based on the information output from the learning model;
An information processing method comprising: a computer executing a process of outputting the generated dialysis prescription. At least one processor;
The processor,
Acquire information on a peritoneal dialysis patient, the information including at least one of edema, urine volume, and water removal volume;
inputting the acquired information into a learning model that has been trained to output information related to a dialysis prescription for a peritoneal dialysis patient when the information of the peritoneal dialysis patient is input, and executing a calculation by the learning model;
generating a dialysis prescription based on the information output from the learning model;
An information processing device that outputs the generated dialysis prescription.

Images