JP7825012B2 - Nursing Information Processing System - Google Patents

Description

The present invention relates to a nursing information processing system.

Japan's population is aging, and the proportion of the population aged 75 and over is expected to continue to increase. As a result, demand for medical care and nursing care is expected to increase further. In light of this situation, there is a shift in the way medical care is provided from the traditional "hospital-based" model to a "community-based" model, in which the entire community supports medical care for people living in their familiar neighborhoods and at home.

In order to promote the shift to community-based medical care, the role of providers of nursing and care at home is important, and there is active development of systems to support these business activities.

For example, Patent Document 1 discloses an invention related to a business management system for home nursing services, which predicts the progression of a patient's condition based on acquired biological information. By predicting progression of the condition, this system contributes to the creation of schedules that assign nurses suited to the patient's condition to provide care at the appropriate time.

Japanese Patent Application Laid-Open No. 2020-021422

In the case of home nursing services specializing in psychiatry, the invention of Patent Document 1 makes it difficult to predict changes in a patient's condition or the worsening of their condition. This is because it is difficult to make a psychiatric diagnosis based solely on biological data information that can be expressed in quantitative terms, such as a patient's pulse, blood pressure, body temperature, and brain waves. In other words, the words a patient speaks and their appearance are important factors in a psychiatric diagnosis.

Therefore, for psychiatric patients, where it is difficult to diagnose or understand changes in symptoms based on biometric data information alone, there is a need for a system that can predict changes in a patient's symptoms based on nursing records that record conversations between nurses and patients and the patient's condition.

In order to solve the above problem, one aspect of the present invention provides a nursing information processing system having a nursing recipient identification information acquisition unit that acquires nursing recipient identification information that identifies a nursing recipient; a nursing recipient identification information storage unit that stores the acquired nursing recipient identification information; a nursing information acquisition unit that acquires nursing information that indicates the content of nursing provided to the nursing recipient and includes text data, in association with the nursing recipient identification information for that nursing; a nursing information storage unit that stores the acquired nursing information; an inference model generation unit that uses the stored nursing information as learning data to generate an inference model for estimating the future condition of the nursing recipient; and an estimation unit that inputs the acquired nursing information into the generated inference model and estimates the future condition of the nursing recipient related to the input nursing information.

According to the present invention, the future condition of a care recipient can be estimated based on nursing information including text data showing the patient's conversations and condition.

A conceptual diagram showing an overview of the nursing information processing system of this embodiment. FIG. 1 is a block diagram showing an example of the functional configuration of a nursing information processing system according to an embodiment of the present invention. Example of hardware configuration for realizing the nursing information processing system of this embodiment A diagram showing an example of the data structure of nursing information. Diagram showing the structure of the training data Flow diagram showing the process for generating training data and generating an inference model Diagram showing the data structure of inference data Flow diagram showing the process of estimation using an inference model Diagram showing the word embedding process An example of outputting hospitalization risk An example of output of hospitalization risk increase/decrease Example of hospitalization risk statistics output An example of output of hospitalization risk statistics trends An example of outputting estimation results in relation to the time of estimation. An example of outputting the estimation results along with the actual progression of symptoms. FIG. 10 is a diagram showing an example of outputting the actual symptom transition and the symptom transition estimated during that period.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the present invention should not be limited to these embodiments and can be embodied in various forms without departing from the spirit and scope of the present invention.
<Embodiment>
<Overview>

1 is a conceptual diagram showing an overview of the nursing information processing system of this embodiment. Nursing information sent from nursing facility A and nursing facility B is stored in a nursing information DB. An inference model for estimating the future condition of a care recipient is generated using the stored nursing information as learning data. By inputting nursing information sent from nursing facility C into this generated inference model, the future condition (e.g., condition next month) of the care recipient related to this nursing information can be estimated.
<Configuration>

Figure 2 is a block diagram showing an example of the functional configuration of the nursing information processing system 200 of this embodiment. As shown in the figure, it has a care recipient identification information acquisition unit 201, a care recipient identification information storage unit 202, a nursing information acquisition unit 203, a nursing information storage unit 204, an inference model generation unit 205, and an estimation unit 206.

Each functional configuration described below is realized as a combination of hardware and software, and can be implemented using a computer such as a personal computer (PC) or server device.

Figure 3 shows an example of the hardware configuration for implementing the nursing information processing system 300, and is composed of a processor 301, a storage device 302, an input/output interface 303, and a communication interface 304. The storage device 302 is composed of RAM and non-volatile memory (HDD, flash memory, etc.), and stores driver programs for controlling the hardware, other application programs, data used in executing the programs, and data obtained as a result of the execution. The processor 301 performs arithmetic processing in accordance with the programs. The input/output interface 303 receives instructions from input devices such as a keyboard or touch panel, and outputs information to an output device such as a display. The communication interface 304 is configured to send and receive data via a network such as the Internet. Each functional component of this embodiment may also be implemented by dedicated hardware.

Furthermore, the present invention can be realized not only as an apparatus, but also as a method. Furthermore, a portion of such an invention can be configured as software. Furthermore, programs used to cause a computer to execute such software, and recording media on which such programs are fixed, are naturally included within the technical scope of this invention (the same applies throughout this specification).

The "care recipient identification information acquisition unit" 201 has the function of acquiring care recipient identification information that identifies the care recipient. The care recipient identification information includes, for example, a "care recipient ID," which is unique identification information, and information about the care recipient identified by this care recipient ID, such as the "name," "gender," "date of birth," "nursing facility," and "medical care expense category." If the care recipient is employed or commutes to a facility, the information may also include information such as the "name of employer" and "name of facility." The functions of this unit are realized by the processor 301, storage device 302, input/output interface 303, etc. It may also be configured to acquire information output from client devices, such as PCs at each facility, as shown in Figure 1, via the communication interface 304.

The "care recipient identification information storage unit" 202 has the function of storing the acquired care recipient identification information, and is realized by creating a database using a storage device 302, etc.

The "nursing information acquisition unit" 203 has the function of acquiring nursing information, including text data, indicating the content of nursing provided to a care recipient, in association with the care recipient identification information for that nursing. Nursing information indicates the content of nursing provided to a care recipient, and is information created for each care recipient ID at the time of nursing. The following description assumes that the nursing in this embodiment is home nursing. Specific examples of nursing information will be provided along with a description of the nursing information storage unit. The functions of this unit are realized by the processor 301, storage device 302, input/output interface 303, etc. It may also be configured to acquire data output from a client device, etc. via the communication interface 304.

Nursing information can also be configured to include medication information. Medication information is information regarding the type of medication prescribed for the care recipient, the dosage, and the method of administration.

The "nursing information storage unit" 204 has the function of storing acquired nursing information, and is realized as a database using the storage device 302 or the like. Figure 4 is a diagram showing an example of the data structure of the stored nursing information. As shown in the figure, nursing information is stored in association with "basic visit information," "basic ADL (Activities of Daily Living) information," "vital signs," and "nursing content" roughly categorized for each care recipient identification ID that uniquely identifies the care recipient.

The "basic visit information" may include, for example, the name of the care recipient, age at the time of the visit, medical care expense category, place of business, start date and time of the visit, end date and time of the visit, absence flag indicating the absence, length of stay, special instruction visit flag indicating that the visit was a special instruction visit, 5 visits per week after discharge flag indicating that the visit was within 5 weeks of discharge, multiple visit flag indicating that multiple visits were made in one day, and multiple visit flag indicating that the visit was made by multiple people.

"Basic ADL information" includes, for example, eating, sleeping, taking medication, self-care, environmental maintenance, financial management, interpersonal relationships, meals, transportation, excretion, changing clothes, bathing, and grooming, while ADL notes include eating (notes), sleeping (notes), taking medication (notes), self-care (notes), environmental maintenance (notes), financial management (notes), interpersonal relationships (notes), eating (notes), transportation (notes), excretion (notes), changing clothes (notes), bathing (notes), and grooming (notes). Here, the above (notes) may include text data.

"Vital signs" include, for example, body temperature, pulse rate, systolic blood pressure, and diastolic blood pressure.

"Nursing content" may include, for example, attentive listening, coordination with the patient's doctor and related institutions, medication and injections, skin and pressure ulcer treatment, assistance with medical examinations, guidance and consultation with family members, rehabilitation, advice, nursing content notes (text), symptom changes, symptom content (text), and special notes (text). The above items marked "(text)" may include text data. Additionally, the medication information described above may be stored as part of the nursing information, along with basic visit information and nursing content.

The "inference model generation unit" 205 has the function of generating an inference model for estimating the future condition of a care recipient using the stored nursing information as learning data. The learning data is generated by vectorizing the text data contained in the nursing information, and is used to generate an inference model for estimating the future condition of a care recipient. The functions of this unit are realized by the processor 301, storage device 302, communication interface 304, etc.

Figure 5 is a diagram showing the structure of the learning data. As shown in Figure 5, the learning data includes, for each care recipient ID, visit information, ADL information, vital signs, nursing content, month of visit, and vector text, which is vectorized text. In addition, if the nursing information includes medication information, the learning data also includes medication information.

Visit information may include, for example, age at time of visit, length of stay, number of days since last visit, number of absences (past 14 days), change in symptoms, special instruction visit flag, 5 visits per week after discharge flag, multiple visit flag, and multiple visits flag.

Examples of ADL information include eating (regular), eating (irregular), eating (refusal), eating (overeating), eating (other), sleep (good sleep), sleep (insomnia), sleep (difficulty falling asleep), sleep (early awakening), sleep (awakening during the night), sleep (other), medication (regular), medication (irregular), medication (refusal), medication (failure to take medication), medication (overdosing), medication (other), financial management (self-management), financial management (management by others), financial management (wasting), cleanliness, environmental maintenance, interpersonal relationships, meals, movement, excretion, changing clothes, bathing, and grooming.

Vital signs include, for example, body temperature, pulse rate, systolic blood pressure, and diastolic blood pressure.

Nursing duties include, for example, listening carefully, coordinating with the patient's doctor and related institutions, administering medication and injections, treating skin and pressure ulcers, assisting with medical appointments, providing guidance and consultation to family members, rehabilitation, and giving advice.

Examples of visit months include visit month flag (January) to visit month flag (December) for January to December.

Vectorized text includes data contained in nursing records that indicate the patient's condition or nursing content, such as ADL notes, symptom content, and nursing content notes and special notes.

Figure 6 is a flow diagram showing the process for generating training data and generating an inference model. This process also uses a tokenizer, a transaction file that stores data in progress, and an embedding layer, all of which are stored in the storage device 302.

As shown in FIG. 6, in step 601, the processor 301 acquires and converts data. If the raw data is a spreadsheet, it is converted to CSV (Comma Separated Value) format. If the raw data is an electronic medical record, the data structure of the nursing information is extracted from the database in CSV format. Each piece of generated CSV-format data is stored in the nursing information database.

In step 602, the processor 301 assigns a target variable to each piece of nursing information. For example, if the patient was hospitalized within the month of the visit, the hospitalization flag is set to "1", and if not, the hospitalization flag is set to "0".

In step 603, the processor 301 creates features. Specifically, the processor 301 calculates and adds to the nursing information data one or more variables selected from the following items: length of stay, number of days since the previous visit, and number of visits in the past 14 days when the patient was absent.

Next, the processor 301 deletes nursing information that is missing information in the ADL basic information and symptom change sections from the nursing information database.

Next, the processor 301 uses an encoder to encode the categorical variable items of the ADL basic information: eating, sleeping, medication, and financial management.

Next, if the medical expense category is missing, the processor 301 copies it from other nursing information for the same care recipient ID to complete it.

Next, the processor 301 converts the visit month into a visit month flag and adds it to the feature.

Next, if the vital sign value is abnormal, the processor 301 deletes the nursing information from the nursing information database. Also, if the vital sign is missing, it complements it with a value of "-1."

The processor 301 then supplements and corrects the text data as appropriate.

In step 604, the processor 301 selects nursing information suitable for the learning data from the nursing information database and stores it in the transaction file. For example, the processor 301 selects only nursing information whose medical expense category is "mental." It may also select only nursing information for care recipient IDs with a predetermined number or more of nursing information items.

In step 605, the processor 301 converts the text data contained in the nursing information stored in the transaction file, indicating the condition of the care recipient or the details of nursing care, into vectors using word embedding, and replaces the original text data with the resulting vectors and stores them.

The text data included in the nursing information indicating the patient's condition or nursing content is selected from one or more of the following: notes on ADL, details of symptoms, notes on nursing content, and special notes on nursing content.

Processor 301 performs word embedding using the following procedure. Processor 301 converts sentences into vectors (positional embedding) using a tokenizer. The tokenizer can use the best publicly available open source at the time.

Then, processor 301 converts these vectors into tensors (word embeddings) by passing them through the embedding layer of a neural network. The embedding layer can use the best publicly available open source software at the time.

Processor 301 generates a tensor with, for example, a first-level: sentence sequence, a second-level: numeric sequence of position embedding results, and a third-level: numeric sequence of word embedding results. Processor 301 then calculates the third-level average of this tensor, pools it (average pooling), and converts it into a vector to create an embedding vector. Processor 201 then stores the resulting data in a transaction file.

In step 606, processor 301 performs data sampling and divides the data in the transaction file into training data:test data = 8:2.

In step 607, processor 301 extracts features to be used for model training from the generated training data. Processor 301 extracts at least visit information, vectorized text, and target variables.

In step 608, processor 301 standardizes and scales the extracted features and outputs the resulting data to a database of training data.

In step 609, processor 301 sequentially reads training data from the training data database and trains the model. Specifically, it trains the model, calculates the loss, and saves the model a predetermined number of epochs.

In step 610, the processor 301 saves the model for the epoch with the smallest loss as the trained model (inference model).

Here, the model can use a deep learning model, but if a better model is available at that time, that model can also be used.

The "estimation unit" 206 inputs the acquired nursing information into the generated inference model and realizes the function of estimating the future condition of the care recipient related to the input nursing information. This acquired nursing information is nursing information related to the estimation of the future condition, and is, for example, nursing information output and acquired from the client terminal of a nursing care provider using this system. The functions of this unit are mainly realized by the processor 301, storage device 302, communication interface 304, etc.

Figure 7 shows the data structure of the inference data that is input into the inference model to estimate the future condition of the care recipient.

As shown in Figure 7, the data structure of the inference data stores, for each care recipient ID, visit information, ADL information, vital signs, nursing content, month of visit, and vector text, which is vectorized text. It can also be configured to store medication information.

Visit information may include, for example, age at time of visit, length of stay, number of days since last visit, number of absences (past 14 days), change in symptoms, special instruction visit flag, 5 visits per week after discharge flag, multiple visit flag, and multiple visits flag.

Examples of ADL information include eating (regular), eating (irregular), eating (refusal), eating (overeating), eating (other), sleep (good sleep), sleep (insomnia), sleep (difficulty falling asleep), sleep (early awakening), sleep (awakening during the night), sleep (other), medication (regular), medication (irregular), medication (refusal), medication (failure to take medication), medication (overdosing), medication (other), financial management (self-management), financial management (management by others), financial management (wasting), cleanliness, environmental maintenance, interpersonal relationships, meals, movement, excretion, changing clothes, bathing, and grooming.

Vital signs include, for example, body temperature, pulse rate, systolic blood pressure, and diastolic blood pressure.

Nursing duties include, for example, listening carefully, coordinating with the patient's doctor and related institutions, administering medication and injections, treating skin and pressure ulcers, assisting with medical appointments, providing guidance and consultation to family members, rehabilitation, and giving advice.

Examples of visit months include visit month flag (January) to visit month flag (December) for January to December.

Vectorized text includes data contained in nursing records that indicate the patient's condition or nursing content, such as ADL notes, symptom content, and nursing content notes and special notes.

Figure 8 is a flow diagram showing the process of inference using the generated inference model. As shown, in step 801, data acquisition and conversion of the data to be processed are performed. The method of data acquisition and conversion is the same as the method of step 601 in Figure 6.

In step 802, the processor 301 creates features. The processing method is the same as the processing method in step 603.

In step 803, processor 301 performs word embedding. The processing method is the same as the processing method in step 605.

In step 804, the processor 301 selects features. The processing method is the same as the processing method in step 607.

In step 805, processor 301 performs scaling. The processing method is the same as the processing method in step 608.

In step 806, the processor 301 applies the obtained data to be processed to an inference model to perform inference and obtain the future possibility of hospitalization (hospitalization risk) as the future condition of the care recipient. The inference result, the possibility of hospitalization, can be obtained as a numerical value such as probability (%), or as a level such as high, medium, or low.

The system can also be configured to acquire the degree of symptoms (a numerical representation of the severity of symptoms) based on a specified symptom assessment scale as a measure of the care recipient's future condition. Examples of symptom assessment scales include the Brief Psychiatric Rating Scale (BPRS) and the PANSS (Positive and Negative Symptom Scale).

The future condition of the care recipient estimated by the estimation unit can also be the risk of the care recipient committing a crime or committing an act of violence, or of committing suicide or self-harm. All of these risks can arise depending on the severity and progression of the symptoms of the mental illness. Furthermore, events that occur depending on the type of illness, symptoms, and progression of symptoms are correlated with ADL and the content of the care recipient's statements. For example, when negative symptoms of schizophrenia appear, motivation decreases and caregiving and bathing (all of which are ADL) tend to be neglected. When positive symptoms appear, hallucinations and delusions appear, resulting in the care recipient's attitude and statements. As these symptoms progress, situations such as suicide or assault can occur. By accumulating information about the care recipient, such as ADL information and content of statements, as nursing information and generating an inference model using the accumulated nursing information, it is possible to estimate the risk of the care recipient committing suicide or assault.

In step 807, the processor 301 outputs the inference result, the possibility of hospitalization, to the input/output interface 303. The inference result can also be output to a client device, etc. via the communication interface 304.

Figure 9 shows the flow of word embedding. As shown in Figure 9, processor 301 tokenizes (converts into the smallest unit of a word sequence) the input text using a tokenizer. In Figure 9, each of the separated words indicated by "tokenization" is a token.

The processor 301 then performs position embedding on the obtained tokens and vectorizes the token sequence. In Figure 9, the numbers indicated by "position embedding" are IDs, which are identifiers uniquely assigned to each token.

Next, processor 301 performs word embedding on each vectorized token and converts the vector into a tensor. In Figure 9, the column of numbers indicated by "word embedding" is a vector uniquely assigned to each token.

Next, processor 301 performs average pooling on the third level of the resulting tensor (the numerical sequence resulting from word embedding) to reshape it and obtain an embedding vector. In Figure 9, the numerical sequence indicated by "average pooling" is the embedding vector after reshaping by average pooling. Here, it has been reshaped into 128 features.

The estimation results from the estimation unit can be output in various ways, and are realized by the processor 301 and input/output interface 303. For example, Figure 10 is a diagram showing an example of outputting the risk of hospitalization as a future condition of a care recipient. As shown in the figure, the hospitalization risk (%) and the change from the previous month (%) are displayed for each patient (care recipient).

Figure 11 shows an example of an output format showing an increase or decrease in hospitalization risk, as another output format. As shown in Figure 11, the increase or decrease in hospitalization risk (%) for each patient (care recipient) is displayed along with the hospitalization risk (%).

Figure 12 shows an example of hospitalization risk statistics, another output format. As shown in Figure 12, the degree of hospitalization risk is tallied and displayed as a pie chart.

Figure 13 shows an example of an output format showing the trend of hospitalization risk statistics, as another output format. As shown in Figure 13, the frequency of the degree of hospitalization risk is tallied and the trend is displayed as a bar graph.

The estimation results by the estimation unit can also be output in association with the estimated time, which is the estimated future time. Figure 14 is a diagram showing the estimated results of the care recipient's future condition for each future month. The vertical axis shows the future severity of symptoms from 0 (low) to 100 (high), and the horizontal axis shows the estimated time, with the estimated results for each month from December 2020 to September 2021 shown as a line graph. By outputting the symptoms along with the estimated time in this way, it is possible to prepare treatments and nursing methods to reduce the predicted peak in symptom severity in May 2021, for example.

The output of the estimation results can also be displayed along with the actual progression of the care recipient's symptoms up to the present. As shown in Figure 15, the graph up to March 2021 shows the progression of the care recipient's symptoms up to that point as a solid line, and the estimated progression of the severity of symptoms from April 2021 onwards as a dotted line.

Furthermore, Figure 16 shows the symptom progression shown in Figure 15, showing both the actual symptom progression (practice) up to March 2021 and the previously estimated symptom progression up to March 2021 (dotted line). As shown in the figure, the previous estimation predicted that the symptom severity would exceed "40" in January 2021, but it can be seen that the symptom severity did not actually increase that much.

By configuring the system to be able to compare estimated results with actual results and obtain the difference between the two, it is possible to verify, for example, the impact that the nursing care provided up to that point has had on the care recipient. It can also be useful for verifying the accuracy of inference models.

The output of estimated results in this way can be useful in supporting nurses and nursing facilities. Because the condition and symptom changes of a care recipient cannot be inferred from biological data shown as numerical values, understanding the condition of the care recipient and providing appropriate care is not easy, and this places a heavy burden on nurses. However, by using this nursing information processing system to view estimated results, it can help improve understanding of the care recipient and develop appropriate nursing care plans, thereby providing support to nurses and nursing facilities.

The estimated results can also be useful in educating nurses. Nursing psychiatric patients is no easy task, as it is difficult to diagnose or grasp changes in symptoms using only biological data information, which can be expressed in numerical terms. In reality, the only way to improve nursing skills is through experience. However, this nursing information processing system estimates the symptoms of the care recipient from nursing information and displays the results. This can be used as a teaching material for nursing education, which can be useful in education aimed at improving nursing skills, etc.

The estimated results do not necessarily have to be output to nursing facilities or the like, but may also preferably be output to the family of the cared person. For example, the system can be configured to associate and store the cared person's identification information with identification information identifying the cared person's family, and allow access to the nursing information processing system by family members identified by the associated identification information, allowing viewing of the estimated results within a specified range. For example, the estimated results shown in Figures 10 to 13 relate to multiple cared persons, so it is not preferable for these family members to view them. However, it is preferable to make the future symptom progression of a single cared person shown in Figures 14 to 16 viewable to the cared person's family members after completing a specified authentication process. For family members involved in home care, understanding the estimated results is extremely useful in providing nursing care.

In addition, if the care recipient is employed or attending a facility, it is also preferable to configure the system to output the estimation results to the company that employs the care recipient or the facility that the care recipient attends. The estimation results can be used to understand symptoms and their progression, which will help to provide appropriate care to the care recipient and provide support to the company that employs the care recipient or the facility that the care recipient attends.

The invention related to the nursing information processing system can also be expressed as an invention related to a method for operating a nursing information processing system, which includes the following steps: a nursing recipient identification information acquisition step for acquiring nursing recipient identification information that identifies a nursing recipient; a nursing recipient identification information retention step for retaining the acquired nursing recipient identification information; a nursing information acquisition step for acquiring nursing information, including text data, that indicates the content of nursing provided to the nursing recipient, in association with the nursing recipient identification information for that nursing care; a nursing information retention step for retaining the acquired nursing information; an inference model generation step for generating an inference model using the retained nursing information as learning data to estimate the future condition of the nursing recipient; and an estimation step for inputting the acquired nursing information into the generated inference model and estimating the future condition of the nursing recipient related to the input nursing information.
<Other Embodiments>

Based on the nursing information processing system described above, it can also be configured to acquire nursing information in association with a nurse. That is, the nursing information processing system described above can further include a nurse identification information acquisition unit that acquires nurse identification information that identifies the nurse providing the care, and a nurse identification information storage unit that stores the acquired nurse identification information, and the nursing information acquisition unit can be configured to acquire nursing information in association with the nurse identification information for that care. The nurse identification information includes, for example, a "nurse ID," which is unique identification information, and information about the nurse identified by this nurse ID, such as the "name," "gender," "date of birth," "place of business," and "years of experience."

With the above configuration, a nurse can output, etc., the estimated future condition of all the care recipients under their care, allowing them to easily understand the estimated results for the future condition of the care recipients they are caring for. Furthermore, as mentioned above, it is also possible to obtain the difference between the estimated results and the actual symptoms, which can be useful for determining whether the care provided by that nurse is contributing to an improvement in the care recipient's symptoms, or whether the nurse and care recipient are compatible. Furthermore, by registering and storing nurse identification information in association with information identifying the nursing facility to which the nurse belongs, it is also possible to extract and output nursing information and estimated results for each nursing facility.

Furthermore, based on the nursing information processing system described above, it is also possible to configure it so that nursing information is acquired in association with a medical institution. That is, the nursing information processing system described above can further include a medical institution identification information acquisition unit that acquires medical institution identification information (medical institution ID) that identifies a medical institution (including a doctor), and a medical institution identification information storage unit that stores the medical institution identification information, and the nursing information acquisition unit can be configured to acquire nursing information in association with the medical institution identification information of the medical institution related to the nursing.

By associating nursing information with a medical institution ID, nursing information can be shared between nursing facilities, care recipients (and their families), and medical institutions. Furthermore, for a medical institution identified by a medical institution ID associated with the nursing information, it is possible to output inferred results about care recipients cared for (associated with) that medical institution or a doctor affiliated with that medical institution, which can be useful for verifying and reviewing nursing care plans for care recipients.

Furthermore, based on the nursing information processing system described above, it is also possible to configure the system to acquire nursing information in association with an administrative agency or welfare service provider (hereinafter referred to as "administrative agency, etc."). That is, the nursing information processing system described above can further include an administrative agency, etc. identification information acquisition unit that acquires administrative agency, etc. identification information (administrative agency, etc. ID) that identifies an administrative agency or welfare service provider (hereinafter referred to as "administrative agency, etc."), and an administrative agency, etc. identification information storage unit that stores the acquired administrative agency, etc. identification information, and the nursing information acquisition unit can be configured to acquire nursing information in association with the administrative agency, etc. identification information of the administrative agency, etc. related to the nursing care.

Administrative agencies include the local government office to which the nursing facility or medical institution belongs, the agency that oversees the nursing facility or medical institution, the local government office where the care recipient resides, the facilities the care recipient attends and the agencies that oversee them, and may also include academic societies related to nursing or medical care. Welfare service providers include nursing facilities and caregiving facilities, as well as businesses and associations that provide a variety of support and assistance services such as assistance with independent living, support for community living, employment transition support, and employment retention support.

By associating nursing information with an administrative agency ID, nursing information can be shared among nursing facilities, care recipients (and their families), and administrative agencies involved with care recipients. In addition, it is possible to output inferred results about care recipients for whom the administrative agency is involved in providing support to the administrative agency identified by the administrative agency ID associated with the nursing information, enabling multifaceted support to be provided to care recipients.
<Mental Health Care System>

The nursing information processing system described above can also be applied to a system for providing mental health care to corporate employees, school students, and others. In the nursing information processing system described above, the nursing recipient can be replaced with an employee (employee or student), and the nursing information can include responses to stress check questions. That is, the system can be configured as a nursing information processing system that includes an employee identification information acquisition unit that acquires employee identification information identifying the employee (employee or student), an employee identification information storage unit that stores the acquired employee identification information, a nursing information acquisition unit that acquires nursing information indicating the content of nursing care provided to the employee (including text data) in association with the employee identification information for that nursing care, a nursing information storage unit that stores the acquired nursing information, an inference model generation unit that uses the stored nursing information as learning data to generate an inference model for predicting the future condition of the employee, and an estimation unit that inputs the acquired nursing information into the generated inference model and predicts the future condition of the employee related to the input nursing information, and the nursing information includes responses to stress check questions.

Here, employees, etc. refer to those who work for a company, etc., such as full-time employees, temporary employees, and part-time workers, as well as students attending or enrolled in various schools and universities, whether private or public.

Furthermore, each component of this system is based on the nursing information processing system for nursing care recipients described above. Employee identification information, like nursing care recipient identification information, includes an "employee ID" that uniquely identifies an employee, as well as information about the employee identified by this employee ID, such as their "name," "gender," "date of birth," "company name," "employment status," "school name," "department name," "class name," and "medical care expense category."

In addition, the nursing information in this system is based on the nursing information in the nursing information processing system related to the care recipient, and is arranged into appropriate information for inferring the mental health of employees, etc. For example, the nursing information is configured to include responses from employees, etc. to questions on a questionnaire administered in the stress check system based on the Industrial Safety and Health Act, which was established for the mental health care of workers. For pupils and students, responses are included in the nursing information using a questionnaire that complies with the stress check system.

Then, just like with nursing information processing systems for care recipients, an inference model is generated using nursing information as learning data. This inference model is used to estimate the future condition of employees, etc. For example, it can estimate whether employees, etc. are in a state of high stress, whether their stress state will increase in the future, or what diseases or symptoms they may be at risk of developing. Furthermore, these estimation results can also be used to help improve the mental health of employees, etc.

Furthermore, similar to the nursing information processing system for cared persons, the system can be configured to output the estimation results to the family of the employee, etc., or to output the estimated condition of the employee, etc. in association with the estimated future time. Furthermore, the system can be configured to acquire nursing information in association with caregiver identification information, acquire nursing information in association with medical institution identification information, or acquire nursing information in association with administrative agency, etc. identification information. In this case, caregivers include the employee's supervisor, health managers at the company employing the employee, or schools attended by students, industrial physicians, stress check implementers, teachers, school doctors, and school counselors. The system can also be configured to output the estimation results to medical institutions identified by medical institution identification information or administrative agencies, etc. identified by administrative agency, etc. identification information. The system can also be configured to acquire the difference between the employee, etc.'s condition in the estimation results at the time of estimation and the actual condition of the employee, etc. The estimated future condition of the employee, etc. can be one or a combination of two or more of the following: hospitalization, assault, crime, and suicide.

The system can also be configured to further include an employer output unit that outputs the estimation results to the employer who employs the employee or the school where the student attends. Employers and schools are required to take into consideration the mental health of their employees and students. By outputting the estimation results to employers and schools, they can understand symptoms and predictions of symptoms, enabling them to respond more appropriately.

Similar to the nursing information processing system for care recipients, the system can be configured to acquire nursing information in association with information identifying medical institutions, administrative agencies, etc., and output estimation results to the medical institutions, administrative agencies, etc. identified by the identification information. By doing so, by sharing nursing information and estimation results, including answers to stress check questions, with the above-mentioned institutions, it is possible to provide multifaceted support to employees, employers, students, and schools.
<Effects>

As described above, the nursing information processing system of the present invention can estimate the future condition of a care recipient based on nursing information that includes text data indicating the care recipient's conversations, state of mind, etc.

200: Nursing information processing system 201: Care recipient identification information acquisition unit 202: Care recipient identification information storage unit 203: Nursing information acquisition unit 204: Nursing information storage unit 205: Inference model generation unit 206: Estimation unit
300: Nursing information processing system 301: Processor 302: Storage device 303: Input/output interface 304: Communication interface

Claims (17)

a nursing information acquisition unit that acquires nursing information including text data that indicates the content of nursing care provided to a care recipient with a mental illness;
an estimation unit that inputs the acquired nursing information into an inference model and estimates future risks that may arise depending on the severity and progression of symptoms of mental illness of the care recipient related to the input nursing information;
an output unit that outputs an estimation result that is a result of the estimation by the estimation unit;
and
The nursing information includes ADL information, and the text data includes notes about ADL.
The inference model is generated using learning data based on nursing information accumulated in the past,
A nursing information processing system in which the possible future risks of the care recipient estimated by the estimation unit are one or a combination of two or more of assault, crime, suicide, and hospitalization. An information storage unit for storing information about the family of the care recipient is further provided,
2. The nursing information processing system according to claim 1, wherein the output unit outputs the inference result to a family member of the care recipient related to the inference result based on the information stored in the information storage unit. The nursing information processing system of claim 1 or 2, wherein the output unit outputs the estimated condition of the care recipient in association with the estimated future time. a caregiver identification information acquisition unit that acquires caregiver identification information that identifies a caregiver providing care;
a caregiver identification information storage unit that stores the acquired caregiver identification information,
4. The nursing information processing system according to claim 1, wherein the nursing information acquisition unit acquires the nursing information in association with the nurse identification information for the nursing care. a medical institution identification information acquisition unit that acquires medical institution identification information that identifies a medical institution or a doctor;
a medical institution identification information storage unit that stores the acquired medical institution identification information,
5. The nursing information processing system according to claim 1, wherein the nursing information acquisition unit acquires the nursing information in association with the medical institution identification information of the medical institution related to the nursing. The nursing information processing system of claim 5, wherein the output unit outputs the estimation result to the medical institution based on the medical institution identification information stored in the medical institution identification information storage unit. an administrative agency etc. identification information acquisition unit that acquires administrative agency etc. identification information that identifies an administrative agency or a welfare service provider;
and an administrative agency etc. identification information storage unit that stores the acquired administrative agency etc. identification information,
7. The nursing information processing system according to claim 1, wherein the nursing information acquisition unit acquires the nursing information in association with the administrative agency or other identification information of the administrative agency or welfare service provider related to the nursing. The nursing information processing system of claim 7, wherein the output unit outputs the estimation result to the administrative agency or welfare service provider based on the administrative agency identification information stored in the administrative agency identification information storage unit. The nursing information processing system of any one of claims 3 to 8, further comprising a difference acquisition unit that acquires the difference between the condition of the care recipient in the estimation result at the time of estimation and the actual condition of the care recipient. an answer information acquisition unit that acquires answer information including text data, which is information indicating answers to questions posed to employees or students who are recipients of mental health care;
an estimation unit that inputs the acquired response information into an inference model and estimates future risks that may arise due to the mental health of the employee or student related to the input response information;
The response information includes responses to questions in a stress check,
The answer information includes ADL information, and the text data includes notes about the ADL.
the inference model is generated using learning data based on previously accumulated answer information;
An information processing system in which the possible future risk of the employee or student estimated by the estimation unit is one or a combination of two or more of assault, crime, suicide, and hospitalization. An information storage unit that stores information about an employer that employs the employee or a school that the student attends,
The information processing system according to claim 10, further comprising an employer output unit that outputs the estimation result, which is the result of the estimation performed by the estimation unit based on the information stored in the information storage unit, to the employer who employs the employee or the school where the student attends. a nursing information acquisition step of acquiring nursing information including text data that indicates the content of nursing provided to a care recipient with a mental illness;
an estimation step of inputting the acquired nursing information into an inference model and estimating future risks that may arise depending on the severity and progression of symptoms of the mental illness of the care recipient related to the input nursing information;
and
The nursing information includes at least ADL information, and the text data includes at least notes about ADL,
The inference model is generated using learning data based on nursing information accumulated in the past,
The possible future risk of the care recipient estimated in the estimation step is one or a combination of two or more of assault, crime, suicide, and hospitalization.
A method for operating a nursing information processing system. an answer information acquisition step of acquiring answer information including text data, the answer information indicating answers to questions posed to employees or students who are recipients of mental health care;
An estimation step of inputting the acquired response information into an inference model and estimating future risks that may arise due to the mental health of the employee or student related to the input response information;
and
The response information includes responses to questions in a stress check,
The answer information includes ADL information, and the text data includes notes about the ADL.
the inference model is generated using learning data based on previously accumulated answer information;
The possible future risk of the employee or student estimated in the estimation step is one or a combination of two or more of assault, crime, suicide, and hospitalization.
A method for operating an information processing system. a nursing information acquisition step of acquiring nursing information including text data that indicates the content of nursing provided to a care recipient with a mental illness;
an estimation step of inputting the acquired nursing information into an inference model and estimating future risks that may arise depending on the severity and progression of symptoms of the mental illness of the care recipient related to the input nursing information;
A program including instructions for causing an electronic computer to execute the following:
The nursing information includes at least ADL information, and the text data includes at least notes about ADL,
The inference model is generated using learning data based on nursing information accumulated in the past,
The possible future risk of the care recipient estimated in the estimation step is one or a combination of two or more of assault, crime, suicide, and hospitalization.
program. an answer information acquisition step of acquiring answer information including text data, the answer information indicating answers to questions posed to employees or students who are recipients of mental health care;
An estimation step of inputting the acquired response information into an inference model and estimating future risks that may arise due to the mental health of the employee or student related to the input response information;
A program including instructions for causing an electronic computer to execute the following:
The response information includes responses to questions in a stress check,
The answer information includes ADL information, and the text data includes notes about the ADL.
the inference model is generated using learning data based on previously accumulated answer information;
The possible future risk of the employee or student estimated in the estimation step is one or a combination of two or more of assault, crime, suicide, and hospitalization.
program. A method for generating an inference model using a computer, comprising:
a nursing information acquisition step of acquiring nursing information including text data that indicates the content of nursing provided to a care recipient with a mental illness;
and an inference model generation step of generating an inference model for estimating future risks that may arise depending on the severity and progression of symptoms of the mental illness of the care recipient by having a computer learn the acquired nursing information as learning data,
The nursing information includes at least ADL information, and the text data includes at least notes about ADL,
The possible future risk of the care recipient estimated by the inference model is one or a combination of two or more of assault, crime, suicide, and hospitalization.
How to generate an inference model. A method for generating an inference model using a computer, comprising:
an answer information acquisition step of acquiring answer information including text data, the answer information indicating answers to questions posed to employees or students who are recipients of mental health care;
and an inference model generation step of generating an inference model for estimating future risks that may arise due to the mental health of employees or students by having a computer learn the acquired response information as learning data,
The response information includes responses to questions in a stress check,
The answer information includes ADL information, and the text data includes notes about the ADL.
The possible future risk of the employee or student estimated by the inference model is one or a combination of two or more of assault, crime, suicide, and hospitalization.
How to generate an inference model.