Setting advice device, dialysis device, learning device, and dialysis information system

Abstract

Provided are a setting recommendation device, a dialysis device, a learning device, and a dialysis information system that recommend a most suitable setting in a complex device while reducing the risk of inappropriate treatment. A setting recommendation system (1) that recommends a setting content of a dialysis device (6) used when a patient receives a dialysis treatment includes a learning model obtained by learning an association between patient dialysis condition data and setting data using data of a dialysis treatment DB (5) in which patient dialysis condition data and setting data are stored for each of a plurality of patients who have received a dialysis treatment in the past, the patient dialysis condition data including transition data indicating a transition of various states of the patient in at least one dialysis treatment, the setting data being data of a setting content of the dialysis device set by a medical staff while referring to the patient dialysis condition data, and outputs, to a recommendation output terminal (4), setting recommendation data that is output data based on the learning model, by applying input data including patient dialysis condition data of a patient for whom a setting is to be recommended in the past to the learning model.

Description

TECHNICAL FIELD

[0001] The present application relates to a setting recommendation device, a dialysis device, a learning device, and a dialysis information system. BACKGROUND

[0002] Conventionally, a dialysis device provided at the bedside of a dialysis patient is used to perform hemodialysis treatment on the dialysis patient. In recent years, along with the sophistication of medical equipment including dialysis devices, the setting of operating parameters for operating the equipment has become complicated, and the setting of the medical equipment requires technical know-how or proficiency. Therefore, a hemodialysis device is disclosed that performs selection of a water removal mode corresponding to individual differences, prediction and adjustment of blood pressure fluctuations in an adjustment unit based on an artificial intelligence (AI) algorithm (for example, refer to Patent Literature 1).

[0003] PRIOR ART DOCUMENTS

[0004] PATENT LITERATURE

[0005] Patent Literature 1: Japanese Patent Application Publication No. 2019-213858 SUMMARY

[0006] PROBLEMS TO BE SOLVED BY THE INVENTION

[0007] Even if an algorithm using artificial intelligence (AI), a statistical method, or the like is used, the device does not necessarily make an appropriate prediction or adjustment. In particular, when an algorithm using artificial intelligence (AI), a statistical method, or the like is used in a device related to the treatment of a patient, such as a medical device, there is a risk of leading to inappropriate treatment.

[0008] In addition, the fact that the treatment setting is different due to many factors such as the patient's physical information on the day of treatment, past experience, and the presence or absence of other diseases also becomes a cause of inappropriate treatment when the device uses treatment derived using an algorithm using artificial intelligence (AI), a statistical method, or the like.

[0009] Therefore, an object of the present application is to provide a setting recommendation device, a dialysis device, a learning device, and a dialysis information system that recommend the most appropriate setting in a device for which the setting is complicated while reducing the risk of inappropriate treatment.

[0010] MEANS FOR SOLVING THE PROBLEMS

[0011] The present application relates to a setting suggestion device that suggests a setting content of a dialysis device used when a patient receives a dialysis treatment, the setting suggestion device including: a learning model that is obtained by learning an association between patient dialysis condition data and setting data set with reference to the patient dialysis condition data, the patient dialysis condition data including passage data indicating a passage of various states of a patient in at least one dialysis treatment, the setting data being data of a setting content of the dialysis device set by a medical staff with reference to the patient dialysis condition data, using data of a dialysis treatment database in which the patient dialysis condition data and the setting data are stored for each of a plurality of patients who have received a dialysis treatment in the past; an input reception unit that receives at least a designation of a patient as a suggestion target; and a data output unit that applies input data including the patient dialysis condition data of the past of the patient as the suggestion target received by the input reception unit to the learning model, and outputs setting suggestion data of output data output based on the learning model to a display section.

[0012] Further, it is preferable that, in the setting suggestion device, the input reception unit further receives a target value related to a dialysis treatment, and the data output unit outputs the setting suggestion data of the output data based on the learning model to the display section by further applying the target value related to the dialysis treatment received by the input reception unit to the learning model as the input data.

[0013] Further, it is preferable that, in the setting suggestion device, the patient dialysis condition data includes at least passage data indicating a blood parameter of a blood state of the patient and a value related to a weight of the patient before and after dialysis, and the setting data includes an operation switching threshold parameter data of the dialysis device and data related to water removal.

[0014] Further, it is preferable that, in the setting suggestion device, the data output unit outputs the output data output by the learning model as the setting suggestion data to the display section.

[0015] Further, it is preferable that, in the setting suggestion device, the data output unit outputs the setting data of the dialysis treatment database approximated to the output data output by the learning model as the setting suggestion data to the display section.

[0016] Further preferably, in the setting suggestion device, there is provided a notification condition storage unit that stores a notification condition, and a notification output unit that outputs, to the display unit, notification information corresponding to the notification condition, in a case where at least one of the patient dialysis condition data and the output data received by the input reception unit satisfies the notification condition stored in the notification condition storage unit.

[0017] Further preferably, in the setting suggestion device, there is provided a setting data reception unit that receives the setting data set by the medical staff, and a relearning unit that causes the learning model to learn using the setting data received by the setting data reception unit and the patient dialysis condition data output to the display unit.

[0018] Further preferably, in the setting suggestion device, there is provided a setting data transmission unit that transmits the setting suggestion data to the dialysis device used in the dialysis treatment of the patient of interest, in a case where the setting suggestion data output to the display unit is confirmed.

[0019] Further, in the setting suggestion device, there can be provided a learning unit that extracts an association between the patient dialysis condition data and the setting data from the dialysis treatment database and causes the learning model to learn.

[0020] Further, the present application relates to a dialysis device including: a learning model learned using data of a dialysis treatment database in which patient dialysis condition data and setting data are stored for each of a plurality of patients who have undergone dialysis treatment in the past, the patient dialysis condition data including transition data indicating a transition of various states of a patient in at least one dialysis treatment, the setting data being data of a setting content of the dialysis device set by a medical staff while referring to the patient dialysis condition data; an input reception unit that receives at least a designation of a patient of interest; and a data output unit that applies input data including the patient dialysis condition data of the patient of interest in the past, received by the input reception unit, to the learning model, and outputs, to a display unit, setting suggestion data based on output data output by the learning model.

[0021] Further, the present application relates to a learning device that is a learning device relating to settings of a dialysis device in a dialysis treatment, including: a patient dialysis condition data acquisition unit that acquires patient dialysis condition data including transition data indicating transitions of various states of a patient in at least one dialysis treatment; a setting data acquisition unit that acquires setting data that is data of contents of settings of the dialysis device set by a medical worker while referring to the patient dialysis condition data; and a learning unit that causes a learning model to learn based on an association between the patient dialysis condition data and the setting data.

[0022] Further, the present application relates to a dialysis information system including the above-described setting suggestion device and the above-described learning device, in which the setting suggestion device includes a model registration unit that causes a learning model of the learning device to be stored in a storage unit.

[0023] Further, in the dialysis information system, the learning device can include a model transmission unit communicably connected to the setting suggestion device and configured to transmit the learning model learned by the learning unit to the setting suggestion device, and the model registration unit can be configured to receive the learning model from the learning device and store the learning model in the storage unit.

[0024] Effects of the Invention

[0025] According to the present application, it is possible to provide a setting suggestion device, a dialysis device, a learning device, and a dialysis information system that suggest the most appropriate settings in a device with complex settings while reducing the risk of inappropriate treatment. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 FIG. 1 is a diagram showing the overall structure of a dialysis information system of the present embodiment.

[0027] Figure 2 FIG. 2 is a diagram showing functional blocks of a setting suggestion system of the present embodiment.

[0028] Figure 3 FIG. 3 is a diagram showing an example of a notification condition storage unit of the setting suggestion system of the present embodiment.

[0029] Figure 4 FIG. 4 is a flowchart showing a learning data storage process of the setting suggestion system of the present embodiment.

[0030] Figure 5 FIG. 5 is a flowchart showing a learning process of the setting suggestion system of the present embodiment.

[0031] Figure 6is a flowchart showing the setting advice processing of the setting advice system of the present embodiment.

[0032] Figure 7 is a flowchart showing the setting advice output processing of the setting advice system of the present embodiment.

[0033] Figure 8 is a diagram showing an example of display in the setting advice system of the present embodiment.

[0034] Figure 9 is a diagram showing an example of display in the setting advice system of the present embodiment.

[0035] Figure 10A is a diagram showing an example of another chart display in the setting advice system of the present embodiment.

[0036] Figure 10B is a diagram showing an example of another chart display in the setting advice system of the present embodiment.

[0037] Figure 11 is a diagram showing an example of a progress chart of patient dialysis condition data of the present embodiment.

[0038] Figure 12A is a diagram for explaining each data used in the learning of the deformation pattern.

[0039] Figure 12B is a diagram for explaining each data used in the learning of the deformation pattern. DETAILED DESCRIPTION

[0040] Hereinafter, an embodiment of the present application will be described with reference to the drawings.

[0041] [System structure of dialysis information system 100]

[0042] Figure 1 is a diagram showing the overall structure of the dialysis information system 100 of the present embodiment.

[0043] Figure 2 is a diagram showing the functional modules of the setting advice system 1 of the present embodiment.

[0044] Figure 3 is a diagram showing an example of the notification condition storage 33 of the setting advice system 1 of the present embodiment.

[0045] As shown in Figure 1 , the dialysis information system 100 of the present embodiment is provided with a data processing server 2, a setting advice output terminal 4, a dialysis treatment DB (database) 5, and a plurality of dialysis devices 6.

[0046] The setting suggestion system 1 (setting suggestion device) consists of a data processing server 2 and a suggestion output terminal 4.

[0047] The dialysis information system 100's setting suggestion system 1 receives patient dialysis status data and setting data related to the dialysis devices 6 from multiple dialysis devices 6, and stores the received data in a dialysis treatment DB5. Then, the setting suggestion system 1 uses the patient dialysis status data and setting data stored in the dialysis treatment DB5 to enable a learning model to learn the relationship between the two. Then, when performing dialysis treatment on a patient using a dialysis device 6, the setting suggestion system 1 uses the learned model to suggest settings for the parameters applied to the dialysis device 6.

[0048] The dialysis information system 100 is, for example, a system used in a medical institution where dialysis treatment using dialysis devices 6 is performed. Dialysis and dialysis treatment, as referred to here, also include the removal of water from the patient's body (dehydration) during the general medical procedure of dialysis. In the dialysis information system 100, a suggestion system 1 stores various data related to dialysis treatment in a medical institution in a dialysis treatment DB5 and utilizes the stored data. The data processing server 2 and the dialysis treatment DB5 are, for example, located in the computer room of the medical institution. Additionally, a suggestion output terminal 4 is, for example, located on the floor where the dialysis devices 6 are installed.

[0049] exist Figure 1 In the example shown, the setting suggestion system 1 is directly connected to the dialysis treatment DB5. Additionally, the setting suggestion system 1 can communicate with multiple dialysis devices 6 via a communication network N.

[0050] The communication network N is the network between the suggestion system 1 and the dialysis device 6, such as a LAN (Local Area Network), an Internet line, a portable terminal communication network, etc. The communication network N is not limited to wired or wireless.

[0051] It should be noted that the setting suggestion system 1 and the dialysis treatment DB5 can be connected via the communication network N.

[0052] Next, the structure of each device will be explained in turn.

[0053] [Settings Suggestion System 1]

[0054] like Figure 1 As shown, the suggestion system 1 consists of a data processing server 2 and a suggestion output terminal 4. The data processing server 2 and the suggestion output terminal 4 can be connected in a way that enables direct communication, but they can also be connected via, for example, a communication network N. Furthermore, in Figure 1In the present embodiment, only one proposal output terminal 4 is shown, but there can be a plurality of them. The proposal output terminal 4 realizes the function of an input / output device of the data processing server 2.

[0055] Figure 2 The illustrated setting proposal system 1 is provided with a control section 20, a storage section 30, a communication IF (interface) section 39, an input section 41, and a display section 42.

[0056] The control section 20 is a central processing device (CPU) that controls the setting proposal system 1. The control section 20 cooperates with the above-described hardware and executes various functions by appropriately reading out and executing an operating system (OS) and application programs stored in the storage section 30.

[0057] The control section 20 is provided with a dialysis device data processing section 21, a learning section 25 (learning unit), and a setting proposal processing section 26.

[0058] The dialysis device data processing section 21 performs processing related to data received from each of a plurality of dialysis devices 6 connected to the setting proposal system 1 in a communicable manner.

[0059] The dialysis device data processing section 21 is provided with a dialysis condition data acquisition section 22 (patient dialysis condition data acquisition unit), a setting data acquisition section 23 (setting data acquisition unit), and a DB storage processing section 24.

[0060] The dialysis condition data acquisition section 22 acquires patient dialysis condition data including progress data representing progress of various states of a patient who has undergone dialysis treatment, by receiving it from the dialysis device 6. Here, the progress data can include time-series data representing progress of blood pressure, venous pressure, amount of water removed, speed of water removed, and the like, in addition to progress data of blood parameters representing blood states such as progress data of a circulating blood volume rate (for example, can be represented by ΔBV (%)) and hematocrit measurement values. In addition, the patient dialysis condition data can include at least one of values related to the weight of the patient such as the weight of the patient before and after dialysis, DW (dry weight), and target amount of water removed, in addition to the progress data, and can further include treatment content, age of the patient, dialysis history, blood test results, and the like. A person who operates the dialysis device 6 (hereinafter, also referred to as an operator) registers information of the patient and the like in the dialysis device 6 before and after and during the dialysis treatment. Here, the treatment content includes, for example, leg lifting treatment accompanying a decrease in blood pressure. In addition, various progress data is registered in the dialysis device 6 when the dialysis treatment is performed. Therefore, the patient dialysis condition data is transmitted by the dialysis device 6, and thus the dialysis condition data acquisition section 22 can acquire the patient dialysis condition data.

[0061] The setting data acquisition section 23 acquires setting data of the dialysis device 6, which is data of the setting contents of the dialysis device set by a skilled medical worker at the time of dialysis treatment, by receiving it from the dialysis device 6. Here, the setting data refers to, for example, operation switching threshold parameter data (hereinafter, also simply referred to as parameter data) of the dialysis device 6, and data related to water removal.

[0062] The parameter data is time series data indicating an alarm line. The parameter data is data used in the control of the dialysis device 6, for example, in a manner such that the dialysis device 6 is controlled in a manner such that the water removal rate is increased by adding a water removal rate specified in accordance with the setting in the case where the rate of change in the circulating blood volume at the time of dialysis exceeds the alarm line indicated by the parameter data. The skilled medical worker sets the parameter data in the dialysis device 6 before the dialysis treatment while observing the past (recent) patient dialysis condition data of the patient to be subjected to the dialysis treatment, and confirming the state of the patient on the day.

[0063] The data related to water removal is, for example, the water removal rate, and can also be the target water removal amount, the water removal time, and the like. Here, the water removal rate is calculated based on the DW set by the skilled medical worker based on the clinical data such as the state of the patient, the cardiothoracic ratio, and the like, the body weight before dialysis, the water removal time, and the past (recent) patient dialysis condition data of the patient to be subjected to the dialysis treatment.

[0064] The setting data is transmitted by the dialysis device 6, and thus the setting data acquisition section 23 can acquire the setting data.

[0065] The DB storage processing section 24 associates the patient dialysis condition data and the setting data acquired from the dialysis device 6, and registers them in the dialysis treatment DB 5. Here, the associated patient dialysis condition data includes at least the transition data used in the setting of the parameter data by the skilled medical worker before the dialysis treatment, the body weight before the dialysis treatment, the DW, and the like, and can also include the treatment contents during the dialysis treatment. In addition, the setting data includes the parameter data set by the skilled medical worker before the dialysis treatment, and the data related to water removal. The association between the patient dialysis condition data and the setting data can be made, for example, based on the patient ID (Identification) identifying the patient, and the date of each data.

[0066] The learning unit 25 inputs and learns the association between the patient dialysis condition data and the setting data stored in the dialysis treatment DB 5 to the learning model. Here, the learning unit 25, for example, inputs the patient dialysis condition data used at the time of setting the parameter data and the like as the input data, inputs the parameter data and the data related to water removal included in the setting data as the training data, inputs the learning model of the LSTM (Long short-term memory) with respect to the time-series data, and inputs the learning model of the DNN (Deep neural network) with respect to the data other than the time-series data, and learns them.

[0067] Note that the learning unit 25 performs learning using the LSTM with respect to the time-series data, but is not limited thereto, and can learn using other learning models.

[0068] As the other learning models, the CNN (Convolutional neural network) and the like used in the classification of images and the like can be considered. It is not data of each period that is treated as numerical data with respect to the time-series data, but the numerical data of the whole or a part of the time-series data is treated as the luminance of an image and considered as a series of images, and thus can be applied.

[0069] Then, the learning unit 25 stores the learning model learned in the model storage unit 32.

[0070] In addition, the learning unit 25 can use all the data stored in the dialysis treatment DB 5, but for example, data in which the dialysis treatment is discontinued in the middle can be excluded. Also, the learning unit 25 can input and learn the learning model by patient after extracting the data by patient, or can input and learn the learning model by attribute after extracting the data by the attribute of the patient. As an example of learning by the attribute of the patient, the learning model can be input and learned by the age (era) of the patient, by the dialysis history of the patient, by the weight (range) of the patient, by gender, or by the treatment mode set in the dialysis device 6.

[0071] The setting suggestion processing unit 26 performs processing of outputting the setting suggestion data related to the setting in the dialysis device 6 for the patient for whom the dialysis treatment will be performed later to the display unit 42.

[0072] The setting suggestion processing unit 26 includes a data reception unit 27, a data output unit 28, and a notification output unit 29.

[0073] The data receiving unit 27 receives at least the designation of the recommended target patient from the input unit 41. In addition, the data receiving unit 27 receives, for example, the target water removal amount (target value involved in the dialysis treatment) in addition to the designation of the recommended target patient. Here, the data receiving unit 27 can receive the data calculated by the setting recommendation processing unit 26 in response to the reception of the body weight before the dialysis treatment and the DW of the recommended target patient as the target water removal amount, for example, as described later. Also, the data receiving unit 27 can receive the blood pressure, the venous pressure, the patient age, and the like.

[0074] The data output unit 28 applies the past patient dialysis condition data of the recommended target patient received by the data receiving unit 27 and the target water removal amount as input data to the learning model stored in the model storage unit 32 and outputs, to the display unit 42, the setting recommendation data based on the output data output by the learning model. Here, the past patient dialysis condition data is, for example, the patient dialysis condition data at the time of the most recent dialysis treatment among the data stored in the dialysis treatment DB 5. The data output unit 28 can output the setting recommendation data together with the past patient dialysis condition data or the past setting data regarding the recommended target patient received by the data receiving unit 27.

[0075] The setting recommendation data can be the output data output by the learning model itself or can be the setting data stored in the dialysis treatment DB 5 that approximates the output data output by the learning model. In addition, the data output unit 28 can output a plurality of setting recommendation data so as to be distinguishable to the display unit 42.

[0076] The notification output unit 29 outputs, to the display unit 42, the notification information in a case where the patient dialysis condition data, the output data output by the data output unit 28 satisfy the notification condition stored in the notification condition storage unit 33. The notification information is information that urges, warns, or reminds with respect to the patient dialysis condition data, the output data.

[0077] The storage unit 30 is a storage area of a hard disk, a semiconductor storage element, or the like that stores programs, data, and the like required for the control unit 20 to execute various processes.

[0078] The storage unit 30 includes a program storage unit 31, a model storage unit 32, and a notification condition storage unit 33.

[0079] The program storage unit 31 is a storage area that stores various programs such as programs for executing various functions of the control unit 20. The program stored in the program storage unit 31 can be a program that differs for each function unit or a plurality of function units of the control unit 20 or can be one program.

[0080] The model storage unit 32 is a storage area that stores the learning model learned by the learning unit 25.

[0081] The notification condition storage section 33 is a storage area that stores notification conditions. For example, as shown in Figure 3 the notification condition storage section 33 is a table in which levels, notification conditions, and notification contents are associated.

[0082] The level indicates the level of the notification content, and is a warning, a reminder, or the like.

[0083] The notification condition is a condition of the patient dialysis status data, output data.

[0084] The notification content is a message corresponding to the notification condition.

[0085] The communication IF section 39 is an interface for performing communication with the dialysis device 6 or the like via a communication network N or the like.

[0086] The input section 41 is, for example, an input device such as an operation button, a controller, or the like.

[0087] The display section 42 is, for example, a display device such as an LCD (Liquid Crystal Display).

[0088] [Dialysis treatment DB 5]

[0089] Figure 1 The dialysis treatment DB 5 shown in FIG. 6 is a database that receives and stores various data that are input to the dialysis device 6 managed by the dialysis information system 100 and output as a processing result by the dialysis device 6. The dialysis treatment DB 5 stores dialysis-related data received from the dialysis device 6 by the setting recommendation system 1. Thus, a large amount of data on dialysis-related data of various patients is stored in the dialysis treatment DB 5. Note that the setting recommendation system 1 performs control of data storage or extraction to the dialysis treatment DB 5.

[0090] [Dialysis device 6]

[0091] The dialysis device 6 is a device that includes at least a device body (console) provided with a blood purification circuit having an arterial blood circuit and a venous blood circuit that constitute a blood circuit for extracorporeal circulation of blood of a patient, a blood purifier having a plurality of connection portions connectable to the blood purification circuit and used for purifying blood extracorporeally circulated through the blood circuit, and various treatment units for blood purification treatment disposed in the blood circuit and the blood purifier.

[0092] The dialysis device 6 is connected to various measuring devices (not shown) and obtains values ​​related to the patient's blood status, such as the rate of change in circulating blood volume, hematocrit, oxygen saturation, and PRR (Plasma Refilling Rate).

[0093] Although not shown, the main body of the dialysis device 6 includes a control unit, a storage unit, a touch panel, and a communication IF unit.

[0094] It should be noted that a computer refers to an information processing device that has a control unit, storage device, etc. The setting suggestion system 1 and the dialysis device 6 are both information processing devices that have a control unit, storage device, etc., and are included in the concept of a computer.

[0095] Next, the processing of the dialysis information system 100 will be explained.

[0096] [Data storage processing]

[0097] First, the processing of data stored in the dialysis device 6 will be explained.

[0098] Figure 4 This is a flowchart illustrating the learning data storage processing of the setting suggestion system 1 in this embodiment.

[0099] The data storage process involves processing the data collected for use in the learning model. Therefore, since the dialysis treatment DB5 already contains a sufficient amount of data to support the learning model's learning process, this processing is unnecessary.

[0100] exist Figure 4 In step S (hereinafter, "step S" will be abbreviated as "S") 11, the dialysis device data processing unit 21 of the suggestion system 1 is set to receive a request for patient dialysis status data from the dialysis device 6. The request may include, for example, the patient ID and the order date. The order date refers to the date on which dialysis treatment was performed; the request may specify, for example, the date of the previous dialysis treatment.

[0101] In S12, the dialysis device data processing unit 21 extracts the patient's dialysis status data corresponding to the request from the dialysis treatment DB5 and sends it to the dialysis device 6.

[0102] The dialysis device 6 displays the received patient dialysis status data. Skilled medical staff can observe the displayed patient dialysis status data while setting up the dialysis device 6.

[0103] In S13, the dialysis device data processing section 21 (setting data acquisition section 23) receives the setting data from the dialysis device 6 after the dialysis treatment using the setting data, which is the content set in the dialysis device 6. The setting is not only made before the start of the dialysis treatment, but also appropriately made as needed during the dialysis treatment.

[0104] In S14, the dialysis device data processing section 21 (DB storage processing section 24) associates the patient dialysis condition data transmitted in the process of S12 with the received setting data, and stores them in the dialysis treatment DB 5.

[0105] In S15, the dialysis device data processing section 21 (dialysis condition data acquisition section 22) receives the patient dialysis condition data from the dialysis device 6. Here, in the dialysis device 6, the data stored in the dialysis device 6 is transmitted to the setting advice system 1 when one dialysis treatment ends. Therefore, in the setting advice system 1, the patient dialysis condition data can be received.

[0106] In S16, the dialysis device data processing section 21 (DB storage processing section 24) causes the received patient dialysis condition data to be stored in the dialysis treatment DB 5. Thereafter, the control section 20 ends the present process.

[0107] Note that the order of the processes of S13 and S15 is not limited. In addition, the processes of S13 and S14, and the processes of S15 and S16 can be performed in parallel at the same time.

[0108] In addition, in the case of a patient who is undergoing dialysis treatment for the first time, there is no data of the patient stored in the dialysis treatment DB 5. In this case, the dialysis device data processing section 21 cannot receive the patient dialysis condition data in the process of S12, but stores the setting data in the dialysis treatment DB 5 in the process of S14, and stores the patient dialysis condition data in the dialysis treatment DB 5 in the processes of S15 and S16.

[0109] Through this process, the setting advice system 1 associates the patient dialysis condition data relating to one dialysis treatment with the setting data relating to the setting of the dialysis device 6 made by the skilled medical staff while referring to the patient dialysis condition data, and stores them in the dialysis treatment DB 5. Therefore, in the dialysis treatment DB 5, the patient dialysis condition data and the setting data relating to one dialysis treatment can be stored in pairs. Then, since the setting advice system 1 stores the data transmitted from the dialysis device 6 in the dialysis treatment DB 5 every time the dialysis treatment is performed, a large amount of data of the patient dialysis condition data and the setting data relating to the setting of the dialysis device 6 made by the skilled medical staff while referring to the patient dialysis condition data can be stored in the dialysis treatment DB 5.

[0110] [Learning Process]

[0111] Next, the learning processing of data stored in the dialysis treatment DB5 is explained.

[0112] Figure 5 This is a flowchart illustrating the learning process of the setting suggestion system 1 in this embodiment.

[0113] After storing a large amount of data in the dialysis treatment DB5, the control unit 20 of the setting suggestion system 1 performs this learning process before proceeding with the application of the setting suggestions to be explained.

[0114] exist Figure 5 In step S21, the learning unit 25 of the setting suggestion system 1 extracts correlation data between patient dialysis status data and setting data from the dialysis treatment DB5. More than one correlation data point is extracted here; preferably, multiple correlation data points are extracted.

[0115] In S22, the learning unit 25 uses the extracted correlation data to enable the learning model to learn.

[0116] In S23, the learning unit 25 determines whether the required number of times of learning has been performed on all the data stored in the dialysis treatment DB5. If the required number of times of learning has been performed on all the data stored in the dialysis treatment DB5 (S23: Yes), the control unit 20 transfers the processing to S24. On the other hand, if the required number of times of learning has not been performed on all the data stored in the dialysis treatment DB5 (S23: No), the control unit 20 transfers the processing to S21.

[0117] In step S24, the learning unit 25 stores the learned model in the model storage unit 32. Afterwards, the control unit 20 terminates the process.

[0118] In this way, the setting suggestion system 1 learns by using the correlation data between the patient dialysis status data and the setting data stored in the dialysis treatment DB5, so that the learning model can learn when given input data, and the learning model outputs suggestions that conform to the settings made by skilled medical personnel.

[0119] [Suggestions on handling related settings]

[0120] Next, the process of using the learning model to make suggestions on the settings of the dialysis device 6 for patients who will undergo dialysis treatment after the learning process will be explained.

[0121] Figure 6 This is a flowchart illustrating the setting suggestion processing of the setting suggestion system 1 in this embodiment.

[0122] Figure 7is a flowchart showing the flow of the setting advice output process of the setting advice system 1 of the present embodiment.

[0123] Figure 8 and Figure 9 is a diagram showing an example of display in the setting advice system 1 of the present embodiment.

[0124] Figure 10A and Figure 10B is a diagram showing an example of another chart display in the setting advice system 1 of the present embodiment.

[0125] In S31 of Figure 6 , the setting advice processing section 26 of the setting advice system 1 outputs, based on the screen display operation of the operator, for example, the patient information input screen 71 shown in Figure 8 Here, the operator refers to a person who operates the dialysis apparatus 6, and can be a non-professional medical worker as before the learning model is used.

[0126] Figure 8 The patient information input screen 71 is a screen for inputting and selecting information related to the patient who is the target of the advice for the dialysis treatment to be performed later. The patient information input screen 71 includes a patient specifying section 72, a date specifying section 73, a this time reservation date 74, a weight specifying section 75, a water removal target section 76, a specified date information output section 77, and buttons 78 and 79.

[0127] The patient specifying section 72 is a region for displaying the patient IDs registered in the dialysis information system 100 and selecting the patient ID of the patient who is the target of the advice. When the patient information input screen 71 is displayed, a plurality of patient IDs are displayed.

[0128] The date specifying section 73 is a region for displaying the past reservation dates and the this time reservation date of the patient specified by the patient specifying section 72 and selecting the this time reservation date. When the patient information input screen 71 is displayed, the date specifying section 73 is blank because no specification in the patient specifying section 72 has been made.

[0129] The weight specifying section 75 is a region for inputting the body weight of the patient who is the target of the advice before dialysis at the body weight and inputting the DW of the patient who is the target of the advice at the DW.

[0130] The water removal target section 76 is a region for outputting the target water removal amount automatically calculated in conjunction with the input of the weight specifying section 75. The target water removal amount automatically output to the water removal target section 76 can be changed by being overlaid.

[0131] The designated date information output unit 77 is an area that takes the previous dialysis booking date selected in the date designation unit 73 as the designated booking date and outputs information obtained from the patient's dialysis status data on the designated booking date. It should be noted that when the patient information input screen 71 is displayed, since the date designation unit 73 has not been performed, the value displayed in the designated date information output unit 77 is blank.

[0132] Button 78 is used to start setting suggestions based on information displayed on the patient information input screen 71 and information that can be obtained from the displayed information. Button 79 is used to clear the information selected and entered on the patient information input screen 71.

[0133] exist Figure 6 In step S32, the setting suggestion processing unit 26 of the setting suggestion system 1 accepts the suggestion target patient by the operator selecting the patient ID of the suggested patient from the patient designation section 72 of the patient information input screen 71. For example, by selecting the patient ID, the setting suggestion processing unit 26 changes the designation section 72a of the patient designation section 72, which contains the selected patient ID, to a selected state.

[0134] In S33, the suggested processing unit 26 is set to retrieve the patient dialysis status data of the accepted patient from the dialysis treatment DB5, and this reservation date, including the current reservation date, is reflected in the date designation unit 73. Thus, the date designation unit 73 displays the patient's past reservation dates and the current reservation date specified by the patient designation unit 72.

[0135] In step S34, the setting suggestion processing unit 26 of the setting suggestion system 1 accepts the reservation date by the operator selecting a reservation date from the date designation section 73 of the patient information input screen 71. Here, the operator usually selects today's date from the reservation dates output to the date designation section 73 as the reservation date. By selecting the reservation date, the setting suggestion processing unit 26 changes the designation section 73a of the date designation section 73, which contains the selected reservation date, to, for example, a selected state.

[0136] In step S35, the setting suggestion processing unit 26 extracts the patient ID and the patient's dialysis status data corresponding to the previous dialysis reservation date (designated reservation date) from the dialysis treatment DB5, and reflects it to the designated date information output unit 77 of the patient information input screen 71. As a result, the designated date information output unit 77, which was previously blank, displays the values ​​of the total water removal volume and water removal rate for the designated patient on the designated reservation date.

[0137] The operator confirms the information of the patient information input screen 71. Then, the operator who confirmed the information of the patient information input screen 71 inputs the pre-dialysis weight and DW of the patient of interest to the weight designation section 75 of the patient information input screen 71, confirms that the target water removal amount is automatically output to the water removal target section 76 by calculation based on the value of the weight designation section 75, and changes it as necessary, and selects the key 78. Thus, in S36, the setting suggestion processing section 26 (the data reception section 27) receives the pre-dialysis weight, DW, and target water removal amount of the patient of interest.

[0138] In S37, the setting suggestion processing section 26 performs the setting suggestion output processing.

[0139] Here, based on Figure 7 The setting suggestion output processing is explained.

[0140] In Figure 7 In S41, the setting suggestion processing section 26 (the data reception section 27, the data output section 28) applies the input data including the patient dialysis condition data of the designated reservation day and the calculated target water removal amount to the learning model to obtain output data.

[0141] In S42, the setting suggestion processing section 26 determines whether there is data that satisfies the notification condition included in the notification condition storage section 33 among the patient dialysis condition data and the output data. In the case where there is data that satisfies the notification condition (S42: YES), the setting suggestion processing section 26 causes the processing to proceed to S43. On the other hand, in the case where there is no data that satisfies the notification condition (S42: NO), the setting suggestion processing section 26 causes the processing to proceed to S44.

[0142] In S43, the setting suggestion processing section 26 extracts the notification content that satisfies the notification condition from the notification condition storage section 33.

[0143] In S44, the setting suggestion processing section 26 generates a graph including the patient dialysis condition data input to the learning model and the designated mode data. Here, as the modes that can be designated, there are a setting selection mode and an original mode. The original data is the output data itself obtained in S42. On the other hand, the setting selection data is the setting data stored in the dialysis treatment DB 5 that approximates the output data obtained in S42 and is actually used in the dialysis treatment. Note that there can also be a MIX mode that includes both the setting selection mode and the original mode.

[0144] In S45, the setting suggestion processing section 26 (the data output section 28, the notification output section 29) outputs the generated graph to the graph output section, and in the case where there is notification content, further outputs the suggestion output screen 81 that outputs the notification content to the message output section to the display section 42. Thereafter, the control section 20 ends the present processing.

[0145] Here, based on Figure 9 The advice output screen 81 is explained.

[0146] Figure 9 The advice output screen 81 is a screen that outputs the setting advice data based on the learning model. The advice output screen 81 includes a patient ID section 82, a this-time information output section 83, a designated day information output section 84, a message output section 85, a mode designation section 86, a chart output section 87, and buttons 88 and 89.

[0147] The patient ID section 82 is the patient ID selected in the patient information input screen 71 Figure 8 ).

[0148] The this-time information output section 83 is a part of the setting advice data (setting data) in the this-time dialysis treatment output by the learning model, and includes the water removal rate.

[0149] The designated day information output section 84 outputs the same content as the designated day information output section 77 of the patient information input screen 71.

[0150] The message output section 85 outputs the extracted notification content in the case where there is data that satisfies the notification condition.

[0151] The mode designation section 86 is a region that designates the mode of the output object in the chart. In the example shown in Figure 9 , the MIX mode is selected.

[0152] The chart output section 87 outputs the chart corresponding to the designation of the mode designation section 86. The chart output section 87 includes a patient dialysis condition data transition chart 87a, a setting advice chart 87b in the original mode, and a setting advice chart 87c in the setting selection mode.

[0153] The button 88 is a button for returning to the patient information input screen 71.

[0154] The button 89 is a button that is output in the case where the setting selection mode or the original mode is designated in the mode designation section 86, and is a button for transmitting the original data of the setting advice chart, that is, the setting advice data, to the dialysis device 6.

[0155] The operator can change the selection of the mode designation section 86 so that the chart output section 87 displays Figure 10A the patient dialysis condition data transition chart 87a output in the setting selection mode and the setting advice chart 87c in the setting selection mode shown in Figure 10BThe illustrated trend chart 87a of the patient dialysis condition data output in the original mode and the setting recommendation chart 87b in the original mode. The operator then judges which of the setting recommendation data to use, or not to use the setting recommendation data. As a result thereof, in the case of using the setting recommendation data, the operator controls the display setting recommendation data to be transmitted to the specified dialysis device 6 by selecting the key 89 of the recommendation output screen 81, by the control section 20 (setting data transmission unit).

[0156] Thus, in the dialysis device 6, even if the setting by the skilled medical staff is not performed, the setting using the setting data based on the setting recommendation data received from the setting recommendation system 1 can be performed, and thus the setting can be easily performed without performing the complicated setting.

[0157] Thus, according to the dialysis information system 100 of the present embodiment, the following effects are exerted.

[0158] In the present embodiment, a learning model is used, which is obtained by learning the correlation between the patient dialysis condition data and the setting data set with reference to the patient dialysis condition data, using the dialysis treatment DB 5 that stores the patient dialysis condition data including the trend data indicating the trend of various states of the patient in at least one dialysis treatment and the setting data of the dialysis device 6 set by the skilled medical staff while referring to the patient dialysis condition data for each of a plurality of patients who have undergone the dialysis treatment in the past. Then, at the time of the dialysis treatment, at least the designation of the recommendation target patient is accepted, the input data including the past patient dialysis condition data of the accepted recommendation target patient is applied to the learning model, and the setting recommendation data based on the output data output from the learning model is output to the display section 42 together with the patient dialysis condition data.

[0159] Thus, at the time of the dialysis treatment, if the patient dialysis condition data of the past of the recommendation target patient is prepared, the setting recommendation data in the dialysis device 6 of the recommendation target patient is output. Therefore, even if the person is not the skilled medical staff, the setting recommendation data can be confirmed, the setting on the dialysis device 6 can be performed, and the setting can be easily performed.

[0160] In addition, in the present embodiment, the output data output from the learning model is output to the display section 42 as the setting recommendation data. Thus, the setting recommendation data itself output from the learning model can be confirmed.

[0161] In addition, in the present embodiment, the setting data stored in the dialysis treatment DB 5 that is similar to the output data is output to the display section 42 as the setting recommendation data. Thus, the setting data of the dialysis treatment having the actual value can be confirmed.

[0162] Also, in the present embodiment, the setting recommendation data of both sides is switched and output or the setting recommendation data of both sides is output at the same time. Thereby, it is possible to confirm while comparing the more appropriate settings for the dialysis device 6.

[0163] Further, in the present embodiment, patient dialysis condition data including transition data indicating a transition of various states of a patient in at least one dialysis treatment and setting data that is data of setting contents of the dialysis device 6 set by a skilled medical staff while referring to the patient dialysis condition data are acquired, and a learning model is learned in association between the acquired patient dialysis condition data and the setting data.

[0164] Thereby, since the setting data set by the skilled medical staff can be learned, the most appropriate setting data with respect to the patient dialysis condition data can be output.

[0165] Note that the present application is not limited to the present embodiment, and variations, modifications, and the like within a range capable of achieving the object of the present application are also included in the present application.

[0166] In the present embodiment, the case where the learning unit is provided in the setting recommendation system is described as an example, but is not limited thereto. The learning unit can be provided in a device different from the setting recommendation system. Then, the learning model learned by the device having the learning unit is transmitted by a model transmission unit of the device (learning device), and thereby the model registration unit of the setting recommendation system acquires and stores the learning model in the storage unit. Thereby, in the setting recommendation system, it is possible to make a recommendation of the setting in the dialysis device 6 in the dialysis treatment of each patient.

[0167] Alternatively, in a case where the device having the learning unit (learning device) and the setting recommendation system are not connected via a communication network or the like, the learning model learned by the device having the learning unit (learning device) can be stored in a storage medium such as a USB, and the storage medium can be used to store the learning model in the storage unit of the setting recommendation system.

[0168] Further, in the present embodiment, the case where the setting recommendation system performs the process of storing data in the dialysis treatment DB 5 is described as an example, but is not limited thereto. Another device can perform the process of storing data in the dialysis treatment DB 5, or each dialysis device 6 can perform the process of storing data in the dialysis treatment DB 5.

[0169] Further, in the present embodiment, the case where the processing result of the setting recommendation processing unit is output to the display unit 42 is described as an example, but is not limited thereto. The device to which the input and output are related can be the dialysis device 6.

[0170] In addition, in the present embodiment, a case where the two screens of the patient information input screen and the suggestion output screen are used is described, but the present application is not limited to this. The operation can be performed on one screen.

[0171] In addition, in the present embodiment, a case where the learning model outputting the setting suggestion data is learned once is described, but the present application is not limited to this. After that, the setting data reception unit of the setting suggestion system can receive the setting data set by the operator based on the setting suggestion data, and the relearning unit of the setting suggestion system can learn the learning model using the received setting data and the patient dialysis condition data output to the display unit at the time of reception or after a certain period of time elapses. Then, the setting suggestion system can update the learning model stored in the model storage unit to the learning model learned by the relearning unit. Thus, the accuracy of the learning model is improved.

[0172] In addition, in the present embodiment, a case where the learning model is learned and one parameter and the water removal rate as the data related to water removal are suggested using the learning model is described, but the present application is not limited to this. For example, a plurality of parameter data such as a parameter for controlling the dialysis device in such a manner that the water removal rate is reduced can be used.

[0173] In addition, in the present embodiment, a case where one setting data is associated with one patient dialysis condition data is described with respect to the association between the patient dialysis condition data and the setting data, but the present application is not limited to this. A case where one setting data is associated with a plurality of patient dialysis condition data can be used in the learning of the learning model.

[0174] Figure 12A and Figure 12B is a diagram for explaining each data used in the learning of the transformation mode.

[0175] Figure 12A A case where the data and the setting data related to the past three dialyses are learned is shown. Here, the past three times are set because it is considered that it is appropriate to include the data up to the third time before one week when a period in which general dialysis treatment is performed is considered.

[0176] Figure 12A is data used at the time of learning, and shows that the learning unit of the setting suggestion system learns the learning model using the association between the setting data of this time and the patient dialysis condition data of the past three times and the target water removal amount of this time. In the case where the learning of this example is performed, as shown in Figure 12B at the time of suggestion, when the setting data of the reservation day of this time is output, the patient dialysis condition data of the most recent three times of the reservation day of this time and the target water removal amount of this time are used as the input data, and are applied to the learning model as the specified reservation day.

[0177] Figure 12A and Figure 12B Examples of the present embodiment consider a case where the setting data is set with reference to a plurality of pieces of patient dialysis condition data.

[0178] In addition, in the present embodiment, a case where the setting data is associated with the most recent patient dialysis condition data is exemplified, but is not limited thereto. The prescribed date can be set to data one week before the present prescribed date, or data within one week from the present prescribed date, for example. In addition, in the present embodiment, the prescribed date is explained in units of days, but can be in units of hours.

[0179] Note that, in the present embodiment and the modified example, a case where the relationship between the input data at the time of learning and the training data, and the relationship between the input data at the time of suggestion and the suggestion data are the same is exemplified, but is not limited thereto.

[0180] In addition, in the present embodiment, a case where the processing of the setting suggestion system different from the dialysis device is exemplified, but is not limited thereto. For example, the dialysis device can have a function of, for example, the setting suggestion processing section as a part of the processing in the setting suggestion system.

[0181] In addition, in the present embodiment, a case where the learning model is caused to learn using all the data stored in the dialysis treatment DB 5 is exemplified, but is not limited thereto. For example, the dialysis treatment DB 5 can store patient dialysis condition data shown in a progress chart 92 shown by the chart display section 91 of the dialysis treatment DB 5. The progress chart 92 shows a case where, when reaching a point in the range 92a from the start of the dialysis treatment, the patient or the dialysis device 6, or the like has some adverse condition, and the dialysis treatment does not proceed smoothly. In a case where there is no adverse condition and the dialysis treatment can proceed smoothly, a chart such as the progress chart 87a of the patient dialysis condition data is generally obtained. Therefore, the control section of the setting suggestion device can determine whether to cause the learning model to learn the input data at the time of learning, and cause the learning model to learn using only the patient dialysis condition data suitable for the input data, based on the waveform of the chart. Figure 11 Figure 9 Note that, in a case where the most recent patient dialysis condition data is the patient dialysis condition data shown by the progress chart 92, since it is data that satisfies the notification condition stored in the notification condition storage section 33, the setting suggestion device outputs a warning (or a reminder). With this notification function, an effect of reducing the risk of inappropriate treatment can be expected.

[0182] Note that, in a case where the most recent patient dialysis condition data is the patient dialysis condition data shown by the progress chart 92, since it is data that satisfies the notification condition stored in the notification condition storage section 33, the setting suggestion device outputs a warning (or a reminder). With this notification function, an effect of reducing the risk of inappropriate treatment can be expected.

[0183] Explanation of Reference Numerals

[0184] 1 setting suggestion system (setting suggestion device) ​

[0185] 2 data processing server

[0186] 4 recommendation output terminal

[0187] 5 dialysis treatment DB

[0188] 6 dialysis device

[0189] 20 control section

[0190] 21 dialysis device data processing section

[0191] 22 dialysis condition data acquisition section

[0192] 23 setting data acquisition section

[0193] 24 DB storage processing section

[0194] 25 learning section

[0195] 26 setting recommendation processing section

[0196] 27 data reception section

[0197] 28 data output section

[0198] 29 notification output section

[0199] 30 storage section

[0200] 31 program storage section

[0201] 32 model storage section

[0202] 33 notification condition storage section

[0203] 41 input section

[0204] 42 display section

[0205] 100 dialysis information system

Claims

1. A setting recommendation device that makes a recommendation on a setting content of a dialysis device used when a patient receives a dialysis treatment, wherein, The setting suggestion device includes: a learning model obtained by learning an association between patient dialysis condition data and setting data set with reference to the patient dialysis condition data, the patient dialysis condition data including passage data indicating a passage of various states of a patient in at least one dialysis treatment, the setting data being data of a setting content of the dialysis device set by a medical staff with reference to the patient dialysis condition data, using data of a dialysis treatment database in which the patient dialysis condition data and the setting data are stored for each of a plurality of patients who have undergone dialysis treatment in the past; an input receiving unit that receives at least a designation of a patient as a suggestion target; and a data output unit that applies input data including the patient dialysis condition data of the patient as the suggestion target received by the input receiving unit to the learning model, and outputs setting suggestion data of output data output based on the learning model to a display section.

2. The setting suggestion device according to claim 1, wherein the input receiving unit further receives a target value related to dialysis treatment, the data output unit outputs the setting suggestion data of the output data output based on the learning model to the display section by further applying the target value related to dialysis treatment received by the input receiving unit to the learning model as the input data.

3. The setting suggestion device according to claim 1 or 2, wherein the patient dialysis condition data includes at least passage data indicating a blood parameter of a blood state of the patient, and a value related to a weight of the patient before and after dialysis, the setting data includes an operation switching threshold value parameter data of the dialysis device and data related to water removal.

4. The setting suggestion device according to claim 1 or 2, wherein the data output unit outputs the output data output by the learning model as the setting suggestion data to the display section.

5. The setting suggestion device according to claim 1 or 2, wherein the data output unit outputs the setting data of the dialysis treatment database approximating the output data output by the learning model as the setting suggestion data to the display section.

6. The setting suggestion device according to claim 1 or 2, comprising: a notification condition storage section that stores a notification condition; and a notification output unit that outputs notification information corresponding to the notification condition to the display section when at least one of the patient dialysis condition data and the output data received by the input receiving unit satisfies the notification condition stored in the notification condition storage section.

7. The setting suggestion device according to claim 1 or 2, comprising: a setting data receiving unit that receives the setting data set by the medical staff; and a relearning unit that causes the learning model to learn using the setting data received by the setting data receiving unit and the patient dialysis condition data output to the display section.

8. The setting proposal device according to claim 1 or 2, comprising: a setting data transmitting unit that transmits the setting proposal data to the dialysis device used in the dialysis treatment of the proposal target patient in a case where the setting proposal data output to the display section is confirmed.

9. The setting proposal device according to claim 1 or 2, comprising: a learning unit that extracts an association between the patient dialysis condition data and the setting data from the dialysis treatment database and causes the learning model to learn.

10. A dialysis device comprising: a learning model that is obtained by causing a learning model to learn an association between patient dialysis condition data and setting data set with reference to the patient dialysis condition data, using data of a dialysis treatment database in which patient dialysis condition data and setting data are stored for each of a plurality of patients who have undergone dialysis treatment in the past, the patient dialysis condition data including passage data indicating a passage of various states of a patient in at least one dialysis treatment, the setting data being data of a setting content of the dialysis device set by a medical staff with reference to the patient dialysis condition data; an input receiving unit that receives at least a designation of a proposal target patient; and a data output unit that applies input data including the patient dialysis condition data of the proposal target patient in the past received by the input receiving unit to the learning model, and outputs setting proposal data based on output data output by the learning model to a display section.

11. Learning device that is a learning device related to a setting of a dialysis device in a dialysis treatment, wherein, The learning device comprises: a patient dialysis condition data acquiring unit that acquires patient dialysis condition data including passage data indicating a passage of various states of a patient in at least one dialysis treatment; a setting data acquiring unit that acquires setting data that is data of a setting content of the dialysis device set by a medical staff with reference to the patient dialysis condition data; and a learning unit that causes a learning model to learn based on an association between the patient dialysis condition data and the setting data.

12. A dialysis information system comprising the setting proposal device according to any one of claims 1 to 8 and the learning device according to claim 11, wherein the setting proposal device comprises a model registering unit that causes the learning model of the learning device to be stored in a storage section.

13. The dialysis information system according to claim 12, wherein the learning device comprises a model transmitting unit that is communicably connected to the setting proposal device and transmits the learning model learned by the learning unit to the setting proposal device, The model registration unit receives the learning model from the learning device and stores it in the storage section.

Images