JP7299054B2 - Assisted Person Management Control Device and Assisted Person Management Control Program - Google Patents

Description

TECHNICAL FIELD The present invention relates to a support recipient management control device, a support recipient management control program, and an AI (artificial intelligence) support device for supporting the work of doctors and nutritionists for support recipients (patients, people requiring nursing care, etc.). be.

Technologies such as Patent Documents 1 to 3 have been disclosed as business support technologies for a support recipient.

In Patent Document 1, a counselor can easily provide appropriate support to individuals who wish to improve their own health by improving their eating behavior and exercise behavior for the prevention of lifestyle-related diseases. A health management support system is described that enables the improvement of living behavior to be achieved reliably and effectively.

This patent document 1 includes means for inputting health management information for each individual from the outside, recording the input information, aggregating the input information, and setting a reference value or / and each individual or each individual input information in advance. Means of comparison and analysis with target values set by individuals and counselors, and at least part of the input information and its aggregate information and comparison/analysis information can be output to specific health care recipients as needed. and means for

In addition, Patent Document 2 describes a technique that can easily realize appropriate life guidance tailored to patients in a small-scale medical institution such as a clinic.

In Patent Literature 2, a lifestyle guidance support device is connected by communication with a plurality of clinic terminal devices and a plurality of patient terminal devices. The lifestyle guidance support device acquires patient information and stores it in the patient information database. The lifestyle guidance support device reads the patient information, transmits it to the external staff terminal device, and receives lifestyle guidance information generated by the external staff terminal device from the external staff terminal device. Lifestyle guidance information is then transmitted from the lifestyle guidance support device to the clinic terminal device. For this reason, life guidance close to direct guidance by professional staff can be realized at a medical institution.

Furthermore, Patent Literature 3 describes a health management system that trend-stores the measured blood glucose level for each user, displays an alarm when the threshold for consecutive days is exceeded, and automatically transmits the measured blood glucose level to a medical institution (medical site). It is

In Patent Document 3, a blood sugar measuring device that can be measured by the subject himself/herself and also has the function of a mobile phone is provided with subject information including the date of measurement, measured blood sugar level data, and mobile phone number/identification information on a medical site. A communication means and a storage means capable of transmitting to a computer are provided, the measured blood sugar level data is stored in the storage means every day along with the measurement date, and an alarm is displayed when the measured blood sugar level data exceeds the threshold for the number of consecutive days, and the medical site and the / Or it is characterized by automatic transmission as trend data of measured blood sugar level to the site of registered dietitian.

By the way, in recent years, the technology of deep learning (a type of deep machine learning), which is one of the logics of AI (artificial intelligence), has rapidly improved and is being used for various purposes.

For example, where people currently visually check the quality of industrial products, images and deep learning have made it possible to detect defective products with higher accuracy than humans can.

Similarly, in the medical field, it is being used for colon examinations, and in the automobile manufacturing field, it is being used for self-driving cars. Deep learning is widely used as a technology in various fields among artificial intelligence.

Deep learning is good at replacing "personalized skills that humans unconsciously pattern and remember by intuition and experience" by using technology to pattern and segment large amounts of data.

Here, the nutrition care work of registered dietitians is also individualized, and the reality is that each person's experience (individual small database) unconsciously creates patterns and conducts nutrition care work.

The nutrition care work flow consists of interviews with the support recipient (patients, people requiring nursing care, etc.), calculations, consideration/problem setting, nutrition care plan creation, implementation, and effect confirmation.

Japanese Patent Application Laid-Open No. 2004-318432 JP 2005-107625 A JP-A-2005-198790

However, registered dietitians have the problem that it takes a lot of time to calculate, consider and set tasks, and create a nutritional care plan in the nutritional care work flow. In addition, there is a problem that the quality of service varies due to individualization.

In this way, in the current nutrition care service, only a small database of individual registered dietitians can be used, so it is becoming difficult for individual registered dietitians to respond appropriately to various support recipients.

In addition, nutrition care work is dependent on the individual and is often done unconsciously (so-called craftsmanship intuition), so training of young people is often inadequate. are different, etc.).

As described above, nutritional care is highly dependent on individual skills, and much of the work is done manually, leading to many problems in service quality and operational efficiency.

An object of the present invention is to obtain a supportee management control device, a supportee management control program, and an AI support device that can improve the quality and work efficiency of nutritional care work.

A support recipient management control device according to the present invention comprises a database storing support recipient information for multiple periods of a plurality of support recipients requiring health management support, and a care plan request for a specific support recipient. a reception unit that accepts, an analysis unit that analyzes changes in the health condition of the specific support recipient received by the reception unit, analysis results of the analysis unit, and support recipient information stored in the database an execution unit for executing care plan analysis processing including setting a problem to be solved by the specific support recipient and creating a care plan for solving the problem, based on and created by the execution unit and a providing unit for providing the care plan to a manager who manages the specific support recipient.

According to the present invention, the reception unit receives a request for a care plan for a particular support recipient. The analysis unit analyzes the transition of the health condition of the specific support recipient received by the reception unit. For example, it analyzes the degree of time-series change in past support recipient information of a specific support recipient stored in the database (for example, the amount and rate of change in body weight or BMI).

Based on the analysis result of the analysis unit and the information of the support recipient stored in the database, the execution unit sets the problem to be solved by the specific support recipient and creates a care plan for solving the problem. Execute the care plan analysis process including

The provision unit provides the care plan created by the execution unit to a manager who manages a specific support recipient.

Conventionally, the above process was dependent on, for example, a manager who manages the health condition of the support recipients, so the more the number of support recipients who are in charge of care, the greater the work burden on the manager. Become. The present invention substitutes for the work performed by this manager, and can improve the quality and work efficiency of nutritional care services.

In the present invention, the manager includes at least a nutrition manager who manages the nutritional aspects of the support recipient, and a medical worker who treats or cares for the disease of the support recipient, and provides support to the manager. and providing a care plan for managing the health condition of the support recipient.

Managers are generally those who help manage the health condition of the recipient (family members, close relatives, etc.), but particularly include nutritionists and medical personnel. Nutrition managers play an important role in managing the nutritional aspects of the recipients of support, and medical staff play an important role in treating or nursing the illnesses of the recipients of support, and have the highest priority as administrators. obtain.

In the present invention, the support recipient uses a menu providing service that provides a daily menu based on the health condition of the support recipient, or provides a meal cooked based on the menu. The nutritional chart information of the support recipient, including the user, presented by a nutrition manager for the menu providing service is stored in the database as the support recipient information.

In the menu providing service, nutrition chart information is essential to obtain the health condition of the support recipient. This nutrition chart information can be diverted as support recipient information of the present invention.

In the present invention, the execution unit is characterized by learning using artificial intelligence to execute analysis for task setting and care plan creation.

By using artificial intelligence to learn analysis for task setting and care plan creation, it is possible to perform analysis in a shorter time than manual analysis. In addition, by using artificial intelligence, as the amount of information of the support recipient stored in the database, the amount of analysis for setting tasks and creating care plans increases, compared to general analysis programs, The reliability of analysis results can be increased.

A supportee management and control program according to the present invention is a supportee management and control program that causes a computer to: It is characterized by operating as

An AI support device according to the present invention has an interactive interface for sending questions about health conditions from a specific support recipient and receiving answers to the questions, and is stored in a database. A care plan providing unit that uses artificial intelligence to provide a care plan to the specific support recipient based on the support recipient information for multiple times of a plurality of support recipients who need the support of and the answers. and have

According to the present invention, an interface interactively asks a question about a health condition and receives an answer to the question from a particular support recipient.

In the care plan providing unit, based on the support recipient information of a plurality of support recipients requiring health management support at multiple times and the answers stored in the database, using artificial intelligence, the identification provide care plans to beneficiaries.

On behalf of the manager, it is possible to always face the support recipient, recognize the support recipient's health condition in an interactive manner, and provide an appropriate care plan.

In the present invention, information on the health condition is provided in the care plan analysis process for creating the care plan in response to a request for information provision from an administrator who manages the health condition of the support recipient. there is

The administrator can obtain information about the health condition in the care plan analysis process for creating the care plan, if necessary.

The present invention is characterized in that the database is provided as a part of the AI support device, or provided on a cloud accessible by communication.

The database may be provided as part of the AI support device, or may be provided on a cloud accessible by communication.

When the database is provided as part of the AI support device, it is not always necessary to have a network communication environment in place, but it is necessary to share information among a plurality of AI support devices.

On the other hand, if it is provided on a cloud that can be accessed by communication, it is not necessary to share information with other AI support devices, but the network communication environment must always be in place.

That is, since there are both advantages and disadvantages, it may be set based on the environment in which the support recipient is placed.

In the present invention, for example, the results of nutritional care (thousands to tens of thousands of cases) performed by many registered dietitians for various supporters are stored in a database on the cloud and aggregated, and this aggregated data It uses deep learning to pattern, segment, and automatically create an appropriate nutritional care plan. As a result, a large amount of data (big data aggregated in a database, not a small database of individuals) can be used to create a standardized and highly reliable nutritional care plan that is not individualized, rather than individually administered by a registered dietitian. It can be produced at high speed and can improve the quality of nutritional care and operational efficiency.

INDUSTRIAL APPLICABILITY As described above, the present invention has the effect of improving the quality and efficiency of nutritional care services.

1 is a schematic diagram of a care plan support system according to this embodiment; FIG. FIG. 10 is a control block diagram showing the flow of processing for creating a care plan using nutritional chart information by executing an automatic menu creation program; 4 is a control flowchart that is a main routine of an automatic menu creation program of menu provision management function A according to the present embodiment. FIG. 3 is a block diagram showing the operations of the care plan management control device 100 according to the present embodiment classified by function. As an example of analysis, it is a conceptual diagram showing an analysis example of each parameter by learning (deep learning) using a neural network. It is a control flowchart for care plan creation control processing based on the execution of care plan support system S according to the present embodiment. It is a schematic diagram of a care plan support system concerning a modification.

FIG. 1 shows a care plan support system S according to this embodiment. The care plan support system S requires big data of the information of the person to be supported, including nutrition chart information, in advance. function based on configuration) is applied. The support recipient information is not limited to the nutritional chart information, and may include other information (problem, care plan, etc. obtained from the nutritional chart information).

FIG. 1 shows a care plan support system S linked with the menu provision management function A. As shown in FIG.

The menu provision management function A provides a menu by the cooperation of the menu provision management business operator 12 and the meal home delivery service business operator 14 at the request of the nutrition manager 10 .

On the other hand, the care plan support system S is mainly composed of the care plan management control device 100, and uses the nutrition chart information obtained from the menu provision management function A to create a care plan for the future of the support recipient. , provides the created care plan to a nutrition manager 10 (doctor, registered dietitian, caregiver, etc.). Below, the created care plan will be described on the premise that it is provided to the nutrition manager 10 .

In addition, when the reliability of care plan creation in the care plan management control device 100 reaches a predetermined level or higher, the created care plan is sent directly to the support recipient (patient, person requiring care, dementia patient, etc.) Although it may be provided (see the modified example), this embodiment will be described on the premise that the care plan created by the care plan management control device 100 is provided to the nutrition manager 10 .

Here, patients, people requiring nursing care, and patients with dementia are particularly exemplified as support recipients. This is because the administrator 10 (for example, registered dietitian) demands more nutritional guidance than other support recipients.

In the care plan support system S of the present embodiment, by sequentially accumulating support recipient information (particularly, in the transitional period, nutrition chart information for judging the success of the care plan is preferable), so-called big data is obtained, and predetermined conditions are satisfied. After that, the big data is used to create a care plan for the support recipient.

In the present embodiment, a predetermined condition as big data is established at the time when the nutritional chart information of 10,000 people and the result observation data for one year or more are accumulated. The establishment of this predetermined condition is an example, and is not limited to the numerical value.

Information used in the above-described menu provision management function A (nutrition chart information, etc.) is used as the transmission source of the supportee information.

In other words, the nutrition manager 10, according to the health condition of the support recipient who ingested the meal prepared by the menu provided based on the nutritional chart information (transition state of weight, BMI, etc. by time-series analysis of the nutritional chart information) Then, a task is set for the person to be supported, and a care plan is created and executed. In addition, success or failure (success or failure) of the care plan is determined by transition of the health condition of the support recipient based on the care plan.

Since this series of work has depended on the nutrition manager 10, the more the number of support recipients in charge of care increases, the more the work load on the nutrition manager 10 increases.

Therefore, in the present embodiment, a care plan support system S is constructed as a substitute for the work performed by the nutrition manager 10 .

On the other hand, the menu provision management function A is applied as initial information (nutrition chart information) for creating big data necessary for the care plan management function in care plan support system S management.

In the following, a configuration in which the care plan management control device 100 and the menu provision management function A are linked will be described. Applicable as an alternative.

(Nutrition chart information source "Menu provision management function A")

First, the menu provision management function A will be described in detail.

As shown in FIG. 1, the care plan support system S according to the present embodiment uses the menu provision management function A to create big data with the nutritional chart information of the support recipient as initial information, and We are building a so-called AI registered dietitian who analyzes data using artificial intelligence (machine learning, deep machine learning "deep learning", etc.) and creates a care plan for the future of the support recipient.

The menu provision management function A of the present embodiment can be used by persons subject to adult diseases and lifestyle-related diseases (including so-called potential patients), children and school children who receive nutritional guidance at kindergartens, nursery schools, schools, etc., and especially nursing care. We do not deny giving nutritional guidance to healthy people who do not need it as support recipients. Therefore, hereinafter, while keeping in mind the person requiring care and the patient with dementia, the term "supported person" will be used, and the patient, the person requiring care, and the patient with dementia will be exemplified as necessary.

The nutrition manager 10 includes a registered dietitian who has the qualifications specified in the Dietitian Law (December 29, 1947 Law No. 245), but the nutrition manager 10 referred to in the present embodiment is a qualified Food coordinators, close relatives (caregivers) who take care of people requiring nursing care or dementia patients at home, and those who manage meals at hospitals and nursing homes This includes those who manage the nutrition of the support recipient and prepare the daily menu, such as the person in charge of the support. In addition, the person who prepares the daily menu may also serve as the cook based on the menu.

In the care plan support system S, the nutrition manager 10 who uses the care plan includes doctors who do not directly engage in menu planning and dietary management.

In the menu provision management function A of the present embodiment, the nutritional value (energy (calorie) and nutrient content) contained in the food is compared with the health condition of the support recipient, etc., and the food is selected and the daily menu is prepared. is designed to create The foods will be briefly described below. The foods to be explained here are mainly foods before cooking and foods after cooking.

Since nutrient balance and total energy (nutrition distribution) are directly related to health management, it is required to devise daily menus for the support recipient.

In particular, when the support recipient is a person requiring nursing care or a patient with dementia, it is important to manage his/her health through daily diet under the control of the nutrition manager 10 .

However, when the nutrition manager 10 considers the daily menu, it is required not only to simply allocate nutrients but also to devise a menu that does not become monotonous.

In addition, it is natural that the nutritional distribution differs depending on the support recipient, and in addition to that, it is necessary to consider the presence or absence of allergies, salt restrictions due to the presence or absence of chronic diseases, preferences, degree of swallowing, number of meals, etc. However, the burden of menu preparation management on the nutrition manager 10 is large.

On the other hand, when outsourcing a meal to an external food delivery service provider, the support recipient or their surrounding collaborators bear the burden of ordering procedures and payment (settlement) of the meal to receive the meal. will be forced.

Therefore, in the present embodiment, based on the automatic menu creation program 18 (see FIG. 2) registered in the menu management control unit 16 of the menu provision management business operator 12, the daily menu necessary for the support recipient is prepared by nutritional management. In addition to automatically creating it on behalf of the supporter 10, cooperation with an external meal delivery service provider 14 is attempted as necessary to provide meals to the supportee.

"Overview of Menu Management Control"

As shown in FIG. 1, each nutritionist 10 has a communication terminal device 20, and the communication terminal device 20 functions as an input source of support recipient information (nutrition chart information). The nutrition chart information includes basic information and nutrition acquisition related information.

The communication terminal device 20 may be a smart phone, a tablet terminal, a personal computer, or the like as long as it can communicate with the menu management control section 16 of the menu providing management business operator 12 via the network 22 .

The basic information in the nutrition chart information of the support recipient is personal information necessary for the support recipient to use the menu provision management function A of the present embodiment.

The basic information in the nutrition chart information of the support recipient includes the support recipient's address, name, e-mail address, password, delivery address when meals are delivered based on the menu, and credit card data. In the case of a person requiring nursing care or a patient with dementia, if the person who pays for the meal delivery is a close relative such as a relative, personal information of the payer such as credit card data is included.

In addition, among the support recipient's nutrition chart information, the nutrition acquisition-related information is condition information necessary for selecting foods when creating a menu.

As an example, the nutrition acquisition-related information includes health management information of the support recipient (including gender, weight, height, blood pressure, pulse rate, and part or all of blood test), amount of exercise per unit period (for example, one day) , Necessary nutritional value per unit period (energy (calorie), nutrient content, etc.), allergy information, chronic diseases (especially diabetes, kidney disease, hypertension, etc. that require dietary restrictions), salt intake limit (e.g., number of grams per day, etc.), preference information, degree of swallowing information, and the number of meals per day and the ratio of nutritional value allocation in each meal.

Table 1 below shows an example of nutritional chart information (basic information and nutrition acquisition related information) of the support recipient. The nutritional chart information shown in Table 1 is the initial information for creating the big data required for the care plan support system S according to the present embodiment, and the large-scale storage device (supported (details will be described later).

なお、栄養取得関連情報の内、健康管理情報は、被支援者を管理する施設（介護施設、学校、医療機関等）に設けられたデータベース（図示省略）にアクセスして、取り込むようにしてもよい。

Among the information related to nutrition acquisition, health management information can also be imported by accessing a database (not shown) provided at a facility that manages the support recipient (nursing facility, school, medical institution, etc.). good.

The communication terminal device 20 can access the menu management control unit 16 of the menu providing management business operator 12 connected to the network 22 by wire or wirelessly via a wireless communication device 24 .

A large-scale storage device 28 is connected to the menu management control unit 16 . Storage areas of the large-scale storage device 28 function as a support recipient information database 28A, a food information database 28B, and a basic menu database 28C. The automatic menu creation program 18 (see FIG. 2) may be stored in a large-scale storage device 28 or a detachable IC memory (not shown) such as a USB memory.

Here, when using the automatic menu creation program 18 (see FIG. 2) of the menu management control unit 16, the menu management control unit 16 cooperates with the automatic menu creation program 18 in response to a request from the communication terminal device 20. Deliver a menu management application program (menu app).

The input of the information of the support recipient in the communication terminal device 20 is executed based on this menu application, and the inputted information of the support recipient is registered in the support recipient information database 28A of the menu management control unit 16 and shared. It has become so.

The food information database 28B stores food information classified by food before cooking and processed food after cooking, and recording the category, classification, nutrient content, and energy of the food.

Information about each food is preferably associated with individual information of the support recipient (nutrition acquisition-related information).

On the other hand, the basic menu database 28C stores menu information provided by the meal home delivery service provider 14, which will be described later, as basic menu information. The menu information is a name by which the completion can be predicted (for example, set meal of mackerel boiled in miso), and includes a photograph of the completion and a recipe as necessary.

The menu information stored in the basic menu database 28C includes nutritional values of seasonings and the like used for cooking in addition to nutritional values of foods.

The menu information is a well-balanced combination of staple foods, side dishes (main dishes and side dishes), dairy products, drinks, sweets, pickles, fruits, etc., and is classified into, for example, breakfast, lunch and dinner, and stored.

The menu information may be a combination of various dishes such as a so-called set meal, or may be a la carte dish (for example, fried chicken, dumplings, etc.). A la carte dishes refer to dishes that have been seasoned with seasonings or the like, or that have undergone heat treatment or the like.

As shown in FIG. 1, a network 22 is connected to a plurality of food delivery service providers 14 affiliated with a menu providing management provider 12 . In this embodiment, there are a plurality of meal delivery service providers 14, but there may be a single meal delivery service provider 14. FIG.

The meal home delivery service provider 14 has a meal preparation department management control unit 30 and a delivery department management control unit 32 . Meal preparation includes a series of tasks such as food procurement, cooking, and presentation.

The meal preparation department management control unit 30 sends menu information of each meal that can be prepared to the menu management control unit 16 via the network 22 . Therefore, the basic menu database 28</b>C of the menu management control unit 16 aggregates the menu information of each meal delivery service provider 14 .

In addition, the meal preparation department management control unit 30 may receive an order for preparation of a meal based on a designated menu from the menu management control unit 16 . Upon receiving an order to prepare a meal based on the menu, the meal preparation department management control unit 30 arranges a worker in charge of cooking and notifies the delivery department management control unit 32 of the delivery destination information.

The delivery department management control unit 32 arranges a delivery worker to deliver the prepared meal to the delivery destination based on the delivery destination information.

Here, in the automatic menu creation program 18 (see FIG. 2) of the menu management control unit 16, the menu information (basic menu information) received from the meal delivery service provider 14 and stored in the basic menu database 28C, and the communication terminal device 20, the individual information of the support recipient (especially information related to nutrition acquisition) is collated, and the nutritional value of energy and nutrients for each meal product (here, nutritional value is a general term for calories and content) ) is analyzed, and menu information for the support recipient (for example, menu information for breakfast, lunch, and dinner for one week) is created.

The procedure for creating the menu information is executed by activating the automatic menu creation program 18, and for example, mathematical programming or machine learning can be applied (see procedures 1 to 4).

(Procedure 1) First, when the automatic menu creation program 18 is executed, the information group 20A (basic information and nutrition acquisition related information) of the support recipient is input from the nutrition manager 10 .

(Step 2) The support recipient's one-week menu is included in the support recipient's nutrition acquisition related information, including health management information, amount of exercise per unit period, required nutritional value per unit period, and daily meals. The basic menu information provided by the meal home delivery service provider 14 and stored in the basic menu database 28C is selected so that the food is not biased based on the number of meals and the allocation ratio of the nutritional value of each meal.

If the food delivery service provider 14 is not affiliated with the food delivery service provider 14, basic menu information may be created by selecting foods from the food information database 28B and stored in the basic menu database 28C.

The number of meals per day is not limited to a maximum of 3 times in the morning, noon and night, and may be 1 or 2 times. Moreover, you may make it include a snack, a midnight snack, and the like. Also, when requesting meal preparation, an upper limit may be set for the cost to be entrusted to the meal delivery service provider 14 .

(Procedure 3) Foods that are not suitable for the support recipient are selected based on allergy information, chronic disease, salt intake limit, preference information, and swallowing degree information from among the support recipient's nutritional acquisition related information, and food information database 28B Replace with similar food based on In addition, if the nutrition manager 10 only wishes to provide the menu, there is no restriction on the food to be replaced. On the other hand, when requesting the meal delivery service provider 14 to prepare a meal based on the menu after replacement, the food to be replaced must be food that can be handled by the meal delivery service provider 14 that provided the basic menu. is preferred. An example of replacement is shown below.

Based on allergen information, select foods that are allergens and replace them with similar foods that are not allergens.

Based on pre-existing conditions, select foods that are specifically marked as aggravating pre-existing conditions and replace them with similar foods that are not specifically defined.

Foods that have a defined total salt intake per serving that is within limits are selected to be eliminated and replaced with similar foods that have total salt intake within limits.

Disliked foods are sorted based on preference information and replaced with similar foods that are not disliked foods.

Based on the swallowing degree information, unmasticable foods are selected and replaced with masticable similar foods.

The first condition for the similar food is that the nutritional value of the food is the same, but it is preferable to select the food to be replaced from the same category.

(Procedure 4) The determined menu information 20B for one week is created and transmitted to the communication terminal device 20 possessed by the nutrition manager 10 .

The nutrition manager 10 who receives the menu information through the communication terminal device 20 prepares meals based on the menu, thereby maintaining the health care of the support recipient.

(Meal delivery service)

On the other hand, if there is a request for delivery of meals from the nutrition manager 10 (supported person), the meal delivery service provider 14 who provides the basic menu information, which is the source of the basic menu information, either collectively for one week or individually The preparation department management control unit 30 is requested to prepare and deliver the meal.

The meal delivery service business operator 14 may only provide menu information to the menu provision management business operator 12 and may not be requested to prepare and deliver meals. , has the advantage of securing new customers.

FIG. 3 is a control flowchart of the main routine of the automatic menu preparation program according to the present embodiment.

In step 50, a process of registering support recipient information (basic information and nutrition acquisition-related information) is executed.

In the next step 52, processing for aggregating the basic menu data is executed. That is, each menu of the meal delivery service provider 14 participating in the menu provision management function A is registered as a basic menu.

In the next step 54, an automatic menu creation process is executed. That is, according to the automatic preparation instruction sent from the communication terminal device 20 possessed by the nutrition manager 10, the menu is automatically prepared based on the pre-registered support recipient information (basic information and nutrition acquisition related information). do.

In the next step 56, meal ordering processing is executed. This meal ordering process is arbitrary, and the nutrition manager 10 (or the support recipient) who has received the menu does not prepare the meal by himself but requests the meal home delivery service provider 14 to prepare and deliver the meal. is executed if

In the next step 58, evaluation reflecting processing is executed. The evaluation is an evaluation after the support recipient has actually eaten a meal prepared based on the received menu. Send to unit 16 .

In the next step 60 , the process (supported person information) up to specification of evaluation points based on the evaluation reflecting process executed in step 58 is transmitted to the care plan management control device 100 . In the care plan management control device 100, this supportee information is sequentially stored in the supportee information database 100A, resulting in big data.

Here, the process up to the identification of the evaluation points is performed manually by the nutrition manager 10, and the support recipient who ingested the meal prepared according to the menu provided based on the nutrition chart information. setting tasks for the support recipient, and creating and executing a care plan based on the health status of the support recipient (weight, BMI, etc. transitional state based on time-series analysis of nutritional chart information). The evaluation score of step 58 is specified based on the success or failure (success or failure) of dealing with the support recipient based on the care plan.

(Support “Care Plan Support System S” based on nutrition chart information)

A care plan support system S according to this embodiment will be described below. The care plan support system S uses the menu provision management function A described above.

As shown in FIGS. 1 and 2, the care plan support system S includes a care plan management control device 100. FIG.

The communication terminal device 20 transmits an information group 20A (basic information and nutrition acquisition related information) of the supportee to the care plan management control device 100 as nutrition chart information (information 1).

The care plan management control device 100 has a large-scale storage device, and its storage area is applied as a support recipient information database 100A.

In the care plan management control device 100, when necessary and sufficient support recipient information is accumulated (big data), the time-series health condition ( weight, BMI, etc.) can be analyzed (calculation of information 2).

In the care plan management control device 100, according to the information 2, artificial intelligence is used to set a problem for a specific support recipient (setting of information 3), and the artificial intelligence is used to solve the problem. create a care plan (a plurality of care plan candidates) for (creation of information 4).

The information 4 is sent to the communication terminal device 20, and in the communication terminal device 20, the nutrition manager 10 can use it as care plan information 20C.

Upon receiving the information 4, the nutrition manager 10 creates care result information actually applied to the support recipient (creation of information 5). The information 5 is sent to the care plan management control device 100 as support recipient information, and is used as a factor for future learning by artificial intelligence.

Information 1 to Information 5 are listed below.
"Information 1" nutrition chart information created by the nutrition manager 10 (see Table 1 above)
"Information 2" Based on big data, analysis information of the time-series health status of specific recipients (undernutrition tendency "weight tendency and BMI tendency" of recipients, risk judgment of malnutrition, etc.) (Table 2) reference)
"Information 3" Task setting by artificial intelligence (deep learning) based on information 2 and the support recipient information stored in the support recipient information database 100A (see Table 3)
"Information 4" Creation of a care plan by artificial intelligence (deep learning) based on information 3 and the supportee information stored in the supportee information database 100A (see Table 4)
"Information 5" Result information after care according to the care plan of information 4 (see Table 5)

なお、栄養カルテ情報に基づくケアプランの抽出の信頼度が所定のレベルに達するまでは、栄養カルテ情報は、ケアが成功と評価される情報に限定することが好ましい。

It is preferable to limit the nutritional chart information to information that evaluates the care to be successful until the reliability of extracting a care plan based on the nutritional chart information reaches a predetermined level.

It is preferable that the plurality of care plan candidates are provided to the nutrition manager 10 as they are, and the nutrition manager 10 side selects, edits, or the like to create a final care plan.

That is, in the case of a skilled nutrition manager 10, it is possible to edit and create an empirically optimal care plan based on a plurality of care plan candidates, thereby reducing the work load of the nutrition manager 10. can do.

On the other hand, in the case of an inexperienced nutrition manager 10, by selecting from among a plurality of care plan candidates, the care plan does not deviate greatly, and by creating a so-called pass line care plan, the experience can be improved. It can cover shallow depths.

FIG. 4 is a block diagram in which the operations of the care plan management control device 100 according to this embodiment are classified by function. Note that each block in FIG. 4 does not limit the hardware configuration of the care plan management control device 100 .

The communication terminal device 20 is connected to the nutrition chart information acquisition section 102 .

The communication terminal device 20 transmits the nutritional chart information shown in Table 1 as the information 1, and the nutritional chart information acquisition unit 102 acquires the information.

The nutritional chart information acquired by the nutritional chart information acquisition unit 102 is sent to the information accumulation processing unit 103 .

The information accumulation processing unit 103 stores the nutritional chart information in the support recipient information database 100A. By executing this storage operation over a long period of time, big data of nutritional chart information is constructed in the supportee information database 100A. For example, by collecting information on nutrition charts of 10,000 people and observation data for one year or more, it can be treated as necessary and sufficient data.

Here, during the period when the care plan management control device 100 continues to collect the nutrition chart information, the communication terminal device 20 operates on both the transmitting side that transmits the nutritional chart information and the user side that uses the care plan support. It has a surface function.

That is, when the communication terminal device 20 transmits the information of the support recipient requesting care plan support (the specific support recipient's nutrition chart information), the nutrition chart information is transferred to the support receiver information database 100A as big data. Along with being stored, the nutritional chart information group reading unit 104 retrieves and reads out the specific support recipient's nutritional chart information.

The nutritional chart information group reading unit 104 is connected to the time-series analysis unit 106 and sends the read nutritional chart information group of the specific support recipient to the time-series analysis unit 106 .

The time-series analysis unit 106 analyzes the health condition of the specific support recipient based on the acquired nutritional chart information group (analysis of information 2).

The analysis result (information 2) in the time-series analysis unit 106 is sent to the malnutrition determination unit 108 .

The malnutrition determination unit 108 determines the malnutrition level based on the analysis result (information 2) of the time-series analysis unit 106 . In this embodiment, the undernutrition level is classified into three types of "small", "medium", and "large", but the classification is not limited to these levels.

The malnutrition determination unit 108 is connected to the output unit 110 and the analysis unit 112 . The determination result (determination level) in the malnutrition determination unit 108 is sent to the communication terminal device 20 via the output unit 110 . In the communication terminal device 20, the determination result is displayed on the monitor or output to a paper medium by a printer, so that the nutrition manager 10 recognizes it.

Here, when the determination level is “medium” or “high” in the malnutrition determination unit 108 , an instruction to create a care plan is output to the analysis unit 112 .

The analysis unit 112 executes analysis processing (problem setting and care plan creation) based on the big data stored in the support recipient information database 100A and the result information (information 2) analyzed by the time-series analysis unit 106. be done.

The analysis unit 112 of the present embodiment is equipped with artificial intelligence, and results information is output for input information by deep learning using the artificial intelligence.

FIG. 5 shows an analysis example of each parameter by learning using a neural network as an example of analysis.

A neural network is a mechanism in which a signal (here, information 2 or information 3) input from an input layer 120 propagates through various nodes 124 (corresponding to neurons) via propagation lines 122 (corresponding to synapses) to an output layer 126. , and deep learning is performed by increasing the number of hidden layers 128 . Result information (here, information 3 or information 4) is output to the output layer 126 .

Analysis unit 112 (see FIG. 4) of the present embodiment executes analysis by deep learning twice in time series.

A first analysis applies information 2 as a factor in the input layer 120 . As a result, the output layer 126 outputs the task as information 3 as the result information.

The second analysis applies the result information (information 3) of the first analysis as input layer 120 factors. As a result, a care plan (a plurality of care plan candidates), which is information 4, is output to the output layer 126 as result information.

Since the first analysis and the second analysis are processed as a single execution pattern, the analysis unit 112 shown in FIG. , the information 4 is created and output.

The analysis unit 112 is connected to the output unit 110 . When the output unit 110 receives the analysis results from the analysis unit 112 (that is, a plurality of care plan candidates as information 4), the output unit 110 sends out the plurality of care plans (information 4) to the communication terminal device 20. FIG. On the communication terminal device 20 side, the plurality of received care plans may be stored and displayed on a monitor, or may be output as a paper medium using an output device such as a printer.

Further, the analysis results (information 3 and information 4) in the analysis unit 112 are fed back to the support recipient information database 100A and stored in association with the support recipient's nutrition chart information.

The operation of this embodiment will be described below according to the flowchart of FIG.

FIG. 6 is a control flowchart for care plan creation control processing based on execution of the care plan support system S. As shown in FIG. Note that this routine is based on the premise that necessary and sufficient big data is secured in the support recipient information database 100A of the care plan management control device 100. FIG.

At step 150, it is determined whether or not the specific support recipient's nutritional chart information (information 1) has been received. If the determination at step 150 is negative, this routine ends.

Further, when the determination in step 150 is affirmative, the process proceeds to step 152 to store the specific support recipient's nutritional chart information in the support recipient information database. This will also lead to the expansion of big data.

In the next step 154, the past to present nutritional chart information of a particular support recipient is read out from the support recipient information database 100A, and then the process proceeds to step 156, time series analysis processing is executed, and information 2 is acquired. and go to step 158 .

In step 158, based on the information 2, the determination of the result of the malnutrition determination is confirmed. In this step 158, when the malnutrition determination is good, it is judged that there is no need to create a care plan, and the process proceeds to step 160 to inform the nutrition manager 10 that the malnutrition level is good (low level). This is notified (transmitted to the communication terminal device 20), and this routine ends.

On the other hand, if the malnutrition determination is unsatisfactory in step 158, it is determined that a care plan needs to be created, and the process proceeds to step 162 to inform the nutrition manager 10 that the malnutrition level is unsatisfactory (medium level or high level). ) is notified (transmitted to the communication terminal device 20 ), and the process proceeds to step 164 .

In step 164, analysis processing (deep learning) for setting the task (information 3) based on big data is executed.

Next, the process proceeds to step 166 to execute analysis processing (deep learning) for creating a care plan (information 4) based on big data, and the process proceeds to step 168 .

In the next step 168 , the information 3 (task) set in step 164 and the information 4 (care plan) created in step 166 are linked to the specific support recipient information 1 received in step 150 . Stored in the support recipient information database 100A. Thereby, the big data of the support recipient information database 100A can be expanded.

In the next step 170, the nutrition manager 10 is notified of the care plan (transmitted to the communication terminal device 20), and this routine ends.

As described above, in the present embodiment, big data of support recipient information is constructed in the support recipient information database 100A of the care plan management control device 100, and nutrition charts of the support recipient registered in chronological order are stored. Based on the information (Information 1), the health status of the support recipient (transitional state of weight, BMI, etc. by time-series analysis of nutritional chart information) is grasped, and analysis processing using big data (deep learning) is performed. I set a task for the support recipient and created a care plan.

Therefore, in the case of a skilled nutrition manager 10, the work load can be reduced, and in the case of an inexperienced nutrition manager 10, the inexperience can be covered.

In the present embodiment, the menu provision management function A is used when constructing the big data that is the premise of the care plan management control device 100 . In the menu provision management function A, it is essential to formulate the process up to the specification of evaluation points manually by the nutrition manager 10. Leave it to S and abolish it.

After a necessary and sufficient amount of data is accumulated in the support recipient information database 100A, the menu provision management function A (menu management control unit 16, large-scale storage device 28) is no longer essential as the care plan support system S. However, by always acquiring the latest nutritional chart information, it is possible to improve the analysis accuracy of the time-series analysis unit 106, so it is preferable to continue the cooperation.

Therefore, in order to improve the care plan creation accuracy of the care plan support system S, after the care plan support system S is introduced, the support recipient information executed in step 60 of FIG. 3 including the evaluation by the automatically created care plan You may continue the transfer process.

Also, the information source that can be big data is not limited to the menu provision management function A, and for example, diagnosis reports of doctors and caregivers may be used.

(Modification)

In the present embodiment, the care plan is provided to the communication terminal device 20 possessed by the so-called human nutrition manager 10. However, as shown in FIG. After the amount of data is accumulated, the AI-equipped device 130 may be connected to the network 22 to function as part of the care plan support system S (terminal device of the care plan management control device 100).

The AI function-equipped device 130 is a so-called user interface represented by an AI speaker 130A, a PC 130B, and the like.

In this embodiment, the care plan created by the care plan management control device 100 is provided for nutrition administrators (physicians, registered dietitians, caregivers, etc.). On the other hand, in the modified example, the AI function-equipped device 130 functions as an AI nutrition manager.

For example, when the supportee declares (utters or inputs) their daily health condition (according to the nutrition acquisition-related information in Table 1) to the AI function-equipped device 130, the AI function-equipped device 130 automatically In cooperation with the care plan management control device 100, based on the care plan provided from the care plan management control device 100, the optimal care plan is advised to the support recipient by voice, text, or the like.

As a result, the support recipient can receive the most suitable care plan for him/herself. In addition, by providing a sensor function to the AI function-equipped device 130, the condition of the support recipient (meal intake condition, complexion, movement, etc.) can be monitored, compared with the provided care plan, and in an emergency, nutrition can be monitored. You can contact the administrator.

The flow of processing by the AI function-equipped device 130 according to the modification is as follows.
(1) The AI-equipped device 130 asks the supportee about necessary items by voice or the like.
(2) Recipients answer questions.
(3) The AI function-equipped device 130 records the answers, cooperates with the care plan management control device 100, acquires the information 2 by time-series analysis, and performs malnutrition determination (determination of quality and level).
(4) If the judgment is "bad", it cooperates with the care plan management control device 100 to set a task, create a care plan, and present it to the support recipient.
(5) The person to be supported executes the care plan, and inputs an evaluation as to whether it was successful or not to the AI function-equipped device 130 at a later date. Note that automatic evaluation may be performed by the sensor function of the AI function-equipped device 130 (detection of the person's complexion, motion state, body temperature, etc.).

By continuing the above (1) to (5) and accumulating successful patterns, the responsiveness and reliability of the artificial intelligence of the AI function-equipped device 130 will be improved.

In the modified example above, the AI function-equipped device 130 is positioned as a terminal device of the care plan management control device 100, but each AI function-equipped device 130 directly accesses the support recipient information database 100A for processing. You can let it stand on its own.

A Menu provision management function S Care plan support system 10 Nutrition manager 12 Menu provision management business operator 14 Meal delivery service provider 20 Communication terminal device 22 Network 24 Wireless communication device 16 Menu management control unit 18 Automatic menu creation program (menu management control program)
28 large-scale storage device 28A support person information database 28B food information database 28C basic menu database 30 meal preparation department management control unit 32 delivery department management control unit 20A information group 20B menu information 20C care plan 100 care plan management control device 100A support recipient Person Information Database (Database)
102 Nutrition chart information acquisition unit (reception unit)
103 Information accumulation processing unit 104 Nutrition chart information group reading unit 106 Time series analysis unit (analysis unit)
108 Malnutrition Determination Department (Analysis Department)
110 Output Unit (Providing Unit)
112 analysis part (execution part)

Claims (7)

Transitions in health conditions of a plurality of support recipients who require health management support over multiple periods and problems to be solved according to the transitions in health conditions as teacher data. is input, the first artificial intelligence learned to output the task, the task for the plurality of support recipients, and a care plan for solving the task as teacher data As a second artificial intelligence learned to output the care plan when the task is input,
a reception unit that receives a request for a care plan for a specific recipient;
Acquiring the support recipient information including information on the health condition of the specific support recipient received by the reception unit at multiple periods, and obtaining the specific support recipient based on the obtained support recipient information at the plurality of periods. an analysis unit that analyzes changes in the health condition of
By inputting the analysis result of the analysis unit to the first artificial intelligence, a problem to be solved by the specific support recipient is obtained, and the obtained problem is input to the second artificial intelligence. an execution unit that acquires a care plan for solving the problem;
a provision unit that provides the care plan acquired by the execution unit to an administrator who manages the specific support recipient;
A supportee management control device having The manager includes at least a nutrition manager who manages the nutritional status of the support recipient, and a medical worker who treats or cares for the disease of the support recipient,
The providing unit provides the administrator with a care plan for managing the health condition of the support recipient.
The support recipient management control device according to claim 1, characterized in that: The support recipient includes a user who uses a menu providing service that provides a daily menu based on the health condition of the support recipient, or provides a meal cooked based on the menu. ,
The analysis unit acquires nutritional chart information of a support recipient presented by a nutrition manager as the support recipient information for the menu providing service.
3. The support recipient management control device according to claim 1 or 2, characterized in that: The analysis unit transmits a question about a health condition, and transmits the answer received by the AI support device having an interactive interface for receiving an answer to the question from the specific support recipient. obtained as information,
The support recipient management control device according to any one of claims 1 to 3. 5. The supportee management control according to claim 4, wherein the providing unit provides the supportee information in response to an information provision request from a manager who manages the health condition of the supportee. Device. A database in which the supporter information for the plurality of periods is stored is provided as part of the AI support device, or is provided on a cloud that can be accessed by communication with the AI support device. 6. The support recipient management control device according to claim 4 or 5, wherein: the computer,
Operate as a receiving unit, an analyzing unit, an executing unit, and a providing unit of the supportee management control device according to any one of claims 1 to 6,
Recipient management control program.

Images