Bed control system

Abstract

Provided is a bed control system that can perform bed control more efficiently and with less burden. 【Solution means】A bed allocation unit 131 including one or more trained models that are generated by machine learning using first teacher data including bed data which is data related to the information of a bed, medical staff data which is data related to the information of medical staff, injury and illness data which is data related to injuries and illnesses, and patient data which is data related to the information of a patient, and second teacher data including data related to one or more beds suitable for allocation to a patient or an attendant, and outputs allocation data which is data related to one or more beds suitable for allocation to a patient or an attendant based on first input data including bed data, medical staff data, injury and illness data, and patient data.

Description

Technical Field

[0001] The present invention relates to a bed control system for allocating hospital beds to patients or attendants.

Background Art

[0002] Bed control involves grasping the usage status of beds, patient admissions and discharges, transfers between rooms, departments, buildings, planned transfers out, the patient's medical condition, the presence or absence of attendants, etc., and efficiently operating the beds so that the beds used by patients or attendants can be secured according to the patient's medical condition and urgency. This is an important task for hospitals, but because various factors need to be considered, the working hours and labor tend to increase, and the burden on hospitals has been extremely large.

[0003] In view of such a situation, in recent years, systems for assisting bed control to perform bed control efficiently have been proposed. For example, Patent Document 1 discloses a technique for displaying information necessary for bed control, such as the status of patients in each bed of a hospital, the operating status of the entire bed, and the usage status of each ward or the entire hospital, in a form that is easy to grasp.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, the technology disclosed in Patent Document 1 is only a technology for assisting and streamlining bed control in a hospital, and since the determination of the final bed allocation (final solution) needs to be made manually, it does not reach the solution of the essential problem. That is, the final solution cannot be obtained immediately.

[0006] The present invention has been made in view of such conventional circumstances. One of the objects of the present invention is to provide a bed control system that can perform bed control in a hospital more efficiently and with a lighter burden. Means for Solving the Problems and Effects of the Invention

[0007] A bed control system according to a first aspect of the present invention is a bed control system that assigns a bed in a hospital to a patient or an attendant, and includes bed data that is data related to the information of the bed, medical staff data that is data related to the information of medical staff, injury / sickness data that is data related to injury / sickness, and patient data that is data related to the information of the patient, which is first teacher data, and second teacher data including data related to one or more beds suitable for assignment to the patient or the attendant, generated by machine learning using the first teacher data, and based on the first input data including the bed data, the medical staff data, the injury / sickness data, and the patient data, it can be configured to include a bed assignment unit composed of one or more trained models that output assignment data, which is data related to one or more beds suitable for assignment to the patient or the attendant.

[0008] According to the above configuration, for example, by using a learned model generated by machine learning the records of past bed controls, it is possible to obtain allocation data in a mode (such as the tendency, custom, and orientation of hospital bed allocation) in accordance with past bed controls. That is, when a user is presented with a single bed based on the allocation data, the user can perform only a confirmation operation such as whether that bed is acceptable, and when multiple beds are presented, the user can perform only a selection operation such as selecting one bed from among them, enabling bed control. Thus, it can be performed more efficiently and with a lower burden. As a result, bed control, which conventionally required a certain level of skill, can be performed relatively easily even by those without such skill. In this way, the characteristic of the bed control system according to the present invention lies in that the final solution can be directly obtained, which is significantly different from the prior art that only supports and improves the efficiency of bed allocation (without obtaining the final solution).

[0009] Also, different from the case where the automation of bed control is mechanically realized by a mathematical model (calculation formula), etc., bed control based on reasons such as the difficulty of verbalization and regularization of "tacit knowledge" in a hospital can also be machine-learned and reproduced.

[0010] The bed control system according to the second aspect of the present invention can be configured such that the bed allocation unit includes a first shared learned model generated by machine learning using data that can be shared among different hospitals among the first teacher data and the second teacher data, and a first dedicated learned model generated by machine learning using data that cannot be shared among different hospitals among the first teacher data and the second teacher data.

[0011] According to the above configuration, since the first shared learned model can be shared among different hospitals, the cost associated with generating the learned model can be suppressed. Further, the first shared learned model does not necessarily have to be generated by machine learning using data that can be shared by all hospitals using the bed control system, and it is understood that this includes cases where it can be shared only among some hospitals. In this case, it is reasonable because data can be shared among similar hospitals, and a higher learning effect can be expected compared to performing machine learning with the data of a single hospital. Further, by applying the first dedicated learned model for each hospital, bed control reflecting unique rules and habits for each hospital that cannot be covered by the first shared learned model can be realized.

[0012] The bed control system according to the third aspect of the present invention can be configured such that the bed allocation unit includes an event learned model that is a learned model including event data which is data related to whether or not it is related to the first teacher data and the first input data, and a normal learned model that is a learned model not including event data in the first teacher data and the first input data.

[0013] According to the above configuration, different bed controls can be realized during normal times and during events. Specifically, a normal learned model used during normal times and an event learned model used during events can be prepared separately in advance, and bed control can be performed based on the normal learned model during normal times (normal mode), and bed control can be performed based on the event learned model during events (event mode). Here, the event learned model can be generated, for example, by machine learning based on triage (determining the treatment priority according to the degree of emergency and severity).

[0014] The bed control system according to the fourth aspect of the present invention can be configured to include a shortage calculation unit that outputs data related to beds or medical staff that are insufficient for treating the injuries and illnesses based on the first input data.

[0015] According to the above configuration, since data related to the insufficient beds or medical staff can be output, for example, based on this data, the number of insufficient beds or medical staff can be conveyed to the user, and the user can easily grasp the shortage of beds or medical staff. Further, the shortage calculation unit can also collect data related to the insufficient beds or medical staff for the treatment of injuries and illnesses in a plurality of hospitals via the bed control system 1, and generate and output organized data. That is, based on the situations of a plurality of hospitals, judgments such as "Hospital A lacks beds, but there are vacant beds in Hospital B and Hospital C, so it is possible to transfer patients", or information provision such as "Hospital D lacks doctors specializing in a, but Hospital E has a doctor specializing in a and who has no scheduled surgeries today, so it is possible to send the doctor for support" can be provided. As a result, for example, in the event of an emergency, a plurality of information can be instantaneously collected, and the acceptance / transfer of patients and requests for support of doctors can be promptly carried out.

[0016] The bed control system according to the fifth aspect of the present invention can be configured to include a second input unit for a user to input second input data including the patient data in the form of text or voice, an input guidance unit for guiding the user to input the information required for the first input data as the second input data, and a language recognition unit including one or more learned models generated by machine learning using third teacher data including a thesaurus (synonyms and related words) of terms, and converting and outputting the second input data in a form usable as the first input data.

[0017] According to the above configuration, the user can input the information required for the first input data in the form of text or voice under the guidance of the input guidance unit, so that this bed control system can be easily used.

[0018] The bed control system according to the sixth aspect of the present invention can be configured such that the language recognition unit includes a second shared learned model generated by machine learning using data that can be shared among different hospitals among the third teacher data, and a second dedicated learned model generated by machine learning using data that cannot be shared among different hospitals among the third teacher data.

[0019] According to the above configuration, since the second shared learned model can be shared among different hospitals, the cost related to the generation of the learned model can be suppressed. Also, the second shared learned model does not necessarily have to be generated by machine learning using data that can be shared by all hospitals using the bed control system, and it is understood that cases where it can be shared only among some hospitals are also included. In this case, it is reasonable because data can be shared among similar hospitals, and a higher learning effect can be expected compared to performing machine learning with the data of one hospital. Also, by applying the second dedicated learned model for each hospital, a bed control system that can handle hospital-specific terms (jargon, abbreviations, etc.) that cannot be covered by the second shared learned model can be realized.

[0020] The bed control system according to the seventh aspect of the present invention includes a display unit that displays a list of a plurality of beds when the bed allocation unit outputs allocation data, which is data related to a plurality of beds suitable for allocation to the patient or the attendant; a selection unit that allows the user to select one bed from among the plurality of beds displayed on the display unit; and a correction unit including one or a plurality of learned models that output correction data, which is data for correcting the allocation data based on the user's habits, generated by machine learning using the data related to the plurality of beds and the data related to the bed selected by the selection unit. The display unit can be configured to display a list of the plurality of beds reflecting the user's habits based on the correction data.

Brief Description of the Drawings

[0021]

Figure 1

Figure 2

Figure 3

Embodiments for Carrying Out the Invention

[0022] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the embodiments shown below are examples for embodying the technical idea of the present invention, and the present invention is not limited to the following. Also, this specification does not in any way specify the members shown in the claims as the members of the embodiments. In particular, the dimensions, materials, shapes, relative arrangements, etc. of the components described in the embodiments are not intended to limit the scope of the present invention only to those, but are merely illustrative examples. Note that the sizes and positional relationships of the members shown in each drawing may be exaggerated for clarity of explanation. Further, in the following description, the same names and reference numerals indicate the same or equivalent members, and detailed descriptions will be omitted as appropriate. Furthermore, each element constituting the present invention may be configured such that a plurality of elements are constituted by the same member and one member serves as a plurality of elements, or conversely, the functions of one member may be shared by a plurality of members and realized. Also, the functions or processes of each element constituting the present invention can be realized by artificial intelligence (AI: Artificial Intelligence) within a possible range. [Bed Control System 1]

[0023] The bed control system 1 according to an embodiment of the present invention will be described with reference to FIGS. 1 to 3.

[0024] The bed control system 1 is a system for controlling hospital beds. As shown in FIG. 1, it is composed of a network 11, a user terminal 12, and a server 13.

[0025] Note that the embodiment of the bed control system 1 is not limited to the foregoing. For example, if the user terminal 12 is configured to have the functions of the server 13 described later, the bed control system 1 can also be realized only by the user terminal 12. 〔Network 11〕

[0026] As shown in FIG. 1, the network 11 is a communication system (ordinary Internet network) for communicably connecting a plurality of user terminals 12 and a server 13, and its embodiment is not particularly limited. 〔User Terminal 12〕

[0027] As shown in FIG. 1, the user terminal 12 is an information terminal owned by a hospital and used by a person (hereinafter referred to as "user US") who performs bed control operations in the hospital. For example, various devices such as a personal computer, a tablet, and a smartphone are applicable. The user terminal 12 does not mean a specific information terminal. When the user US uses a plurality of information terminals, all of them correspond to the user terminal 12.

[0028] Further, the user terminal 12 corresponds to an example of the "second input unit", "display unit", and "selection unit" in the claims, and performs data input such as operations of a keyboard or a touch panel, voice input to a microphone, transmission and reception of data with the server 13 via the network 11, transmission of data to the user US, and the like. 〔Server 13〕

[0029] As shown in FIG. 1, the server 13 is a server managed by the administrator AD and is communicably connected to the user terminal 12 via the network 11.

[0030] Further, as shown in FIG. 2, the server 13 includes a bed allocation unit 131, a shortage calculation unit 132, an input guidance unit 133, a language recognition unit 134, a correction unit 135, and a learning unit 136, and performs processing related to bed control based on the data transmitted via the user terminal 12. 〈Bed Assignment Unit 131〉

[0031] The bed assignment unit 131 is a member composed of a plurality of learned models generated by the learning unit 136 through machine learning using first teacher data including bed data which is data related to bed information, medical staff data which is data related to information of medical staff, injury and illness data which is data related to injury and illness, and patient data which is data related to patient information, and second teacher data including data related to beds suitable for assignment to patients or accompanying persons based on the first teacher data.

[0032] Also, when actually performing bed control, the bed assignment unit 131 outputs assignment data which is data related to one or more beds suitable for assignment to a patient or an accompanying person based on first input data including, at that time, bed data, medical staff data, injury and illness data, and patient data of the patient to whom the bed is actually assigned. 《First Teacher Data》

[0033] The first teacher data is data used for generating the learned models constituting the bed assignment unit 131, and includes bed data, medical staff data, injury and illness data, patient data, and emergency data. That is, when it is determined that a specific patient or accompanying person should be assigned to a specific bed, it is data related to the reasons and causes for that determination. The first teacher data may be data used in past bed control (such as actual patient data), or data created for machine learning of the bed assignment unit 131 (non-existent data).

[0034] The data that can be included in the first teacher data is exemplified below.

[0035] Bed data: Together with the identification information of the bed, data such as the state of the bed, the distance from the nurse station, information on whether it is a large room or a private room, and the state of the ward where the bed is located may be included. Also, data such as the period during which the bed cannot be used due to hospital construction may be included.

[0036] Medical staff data: Along with the identification information of medical staff, it may include data such as what medical staff can do (whether they can perform specific surgeries or not), attendance status, etc.

[0037] Injury and illness data: Along with the identification information of injuries and illnesses, it may include data such as the content of injuries and illnesses, treatment methods (including treatment data such as medication and surgical information), the period required for the treatment, costs, etc.

[0038] Patient data: Along with the identification information of patients, it may include data such as gender, age, economic power, the content of injuries and illnesses, whether they have other underlying diseases, the presence or absence of accompanying persons and the number of them, etc.

[0039] Emergency data: It may include data such as information on whether it is an emergency or normal times, and the content of emergencies (for example, earthquake disasters, pandemics, wartime). Emergency data is prepared in advance based on the knowledge of experts. 《Second teacher data》

[0040] The second teacher data is data used for generating the learned model that constitutes the bed allocation unit 131, and includes the identification information of one or more beds suitable for allocation to patients or accompanying persons when based on the first teacher data.

[0041] Specifically, when the first teacher data is data used in past bed control, the identification information of the actually allocated beds can be used. Also, when the first teacher data is data created for machine learning of the bed allocation unit 131, one or more beds suitable for allocation to patients or accompanying persons are selected based on this, and the identification information of the beds is created.

[0042] That is to say, it can be said that the second teacher data is the judgment (answer of bed control) of the person who performs bed control when assuming the first teacher data. The following cases (examples) can be considered as such judgments.

[0043] Case 1: Since the patient is a disabled infant, assign a bed next to which an attendant's bed can be placed (assign the attendant's bed next to the attendant).

[0044] Case 2: Since the patient is female, assign a bed in which there is no male patient in the same room.

[0045] Case 3: Since the patient makes strange noises, do not assign another bed in the same room to another patient.

[0046] Case 4: During the summer vacation period, since the number of student hospitalizations increases, assign the beds so as to appropriately shorten the length of hospitalization and increase the bed turnover rate.

[0047] Case 5: Since the patient has financial means and desires higher-quality services, assign a bed in a private room (special room).

[0048] Case 6: During the recovery period immediately after surgery, since it is necessary to observe the patient's condition frequently, assign a bed near the nurse station.

[0049] Case 7: Since it is an emergency, assign beds based on triage (determining the treatment priority according to the urgency and severity of the injury or illness).

[0050] Case 8: During the hospitalization period, assign beds so that the movement of beds and rooms is minimized as much as possible.

[0051] As described above, as the judgment of the person in charge of bed control, those such as Cases 1 to 8 can be considered. However, in reality, usually, a plurality of conditions to be considered are intricately intertwined, and the judgment of the person in charge of bed control and the reasons and causes for the judgment do not necessarily take a one-to-one form.

[0052] In such a case, it is considered that the person who controls the bed unconsciously considers various complex conditions simultaneously and further determines (judges) the bed allocation after considering which conditions to prioritize (weighting).

[0053] In the bed control system 1 according to an embodiment of the present invention, such a determination is made by the bed allocation unit 131 generated by machine learning using first teacher data and second teacher data, so that a determination (a determination that is difficult to verbalize or rule out) that is made unconsciously can be mechanized and reproduced. 《First Shared Learned Model 131a and First Dedicated Learned Model 131b》

[0054] In addition, the plurality of learned models included in the bed allocation unit 131 can be classified into a first shared learned model 131a generated by machine learning using data that can be shared among different hospitals among the first teacher data and the second teacher data, and a first dedicated learned model 131b generated by machine learning using data that cannot be shared among different hospitals.

[0055] For example, the first shared learned model 131a can be generated using injury data, emergency data, etc. that can be shared among different hospitals, and the first dedicated learned model 131b can be generated using bed data, medical staff data, patient data, etc. that cannot be shared among different hospitals.

[0056] In addition, the first shared learned model 131a does not necessarily have to be generated by machine learning using data that can be shared by all hospitals using the bed control system 1. For example, it is understood that it includes cases where it can be shared only among some hospitals, such as between hospitals with similar hospital scales (number of beds), between hospitals with similar hospital characteristics (strong in the acute phase, cancer specialty, cardiovascular specialty, organ transplantation specialty, etc.), and between hospitals with similar regional characteristics (regions with many lifestyle diseases because the intake of sugar and salt exceeds the national average, etc.). In this case, it is reasonable because data can be shared between similar hospitals, and a higher learning effect can be expected than performing machine learning with the data of a single hospital.

[0057] In this way, by providing the first shared pre-trained model 131a and the first dedicated pre-trained model 131b independently, the first shared pre-trained model 131a can be shared in different hospitals, so that the cost for generating the pre-trained model can be suppressed. Further, by applying the first dedicated pre-trained model 131b for each hospital, bed control reflecting the unique rules and habits of each hospital can be realized. "Normal pre-trained model 131H and emergency pre-trained model 131U"

[0058] Moreover, the plurality of pre-trained models provided in the bed allocation unit 131 can be classified into a normal pre-trained model 131H used in normal times and an emergency pre-trained model 131U used in emergencies.

[0059] The emergency pre-trained model 131U is different from the normal pre-trained model 131H in that the first teacher data used for its generation includes emergency data and the determination of case 7 is included in the second teacher data.

[0060] In this way, by providing the normal pre-trained model used in normal times and the emergency pre-trained model used in emergencies separately in advance, bed control can be performed based on the normal pre-trained model in normal times (normal mode), and bed control can be performed based on the emergency pre-trained model in emergencies (emergency mode). "First input data"

[0061] The first input data is the data input to the pre-trained model constituting the bed allocation unit 131 when actually performing bed control. As the content of the data, similar to the first teacher data, bed data, medical staff data, injury data, patient data, and emergency data may be included.

[0062] The method of inputting the input data is not particularly limited. For example, among the bed data, medical staff data, injury data, and patient data, those whose data such as gender, age, and economic power are already known in advance can be input in advance, or obtained in cooperation with an external system such as an electronic medical record, and the data that is determined immediately before the allocation of the bed, such as the content of the bed and emergency data, among the patient data, may be input at that timing. 《Allocation Data》

[0063] The allocation data is data output from the learned model that constitutes the bed allocation unit 131, and is data related to one or more beds suitable for allocation to a patient or an attendant, output based on the first input data, and may include data such as the identification information of the bed and the allocation period.

[0064] The output allocation data is transmitted to the user terminal 12 and transmitted to the user US. The method of transmitting to the user US is not particularly limited, and it may be displayed in the form of characters or graphics on the display, or transmitted by voice. For example, ct2-6 to ct2-8 in FIG. 3 are in a state where the identification information of a plurality of beds suitable for allocation to a patient or an attendant is listed in the form of characters. 《Other Aspects of the Bed Allocation Unit 131》

[0065] Note that the aspect of the bed allocation unit 131 is not limited to the foregoing.

[0066] For example, it is not necessary to solely assign the output of the allocation data to the learned model that constitutes the bed allocation unit 131, and a part of it may be shared with other methods such as a mathematical model. For example, the learned model is responsible for the part that is difficult to process mechanically (judgment that is difficult to verbalize or rule-based), and the mathematical model is responsible for the part that can be processed mechanically, so that the allocation data can be output more efficiently.

[0067] In addition, it is not necessary to have all of the learned models for outputting allocation data in the server 13, and a part or all of them may be provided in the user terminal 12. For example, it is also possible to provide the first shared learned model 131a in the server 13 and the first dedicated learned model 131b in the user terminal 12. That is, the determination that can be shared among different hospitals is made in the server 13, and the parts that require unique determination for each hospital are made in the user terminal 12 of each hospital, so that the cost of operating the system can be suppressed.

[0068] In addition, the learned models that make up the bed allocation unit 131 may be provided by separating the learned models for each season in accordance with the prevalence of diseases and injuries for each season.

[0069] In addition, the learned models that make up the bed allocation unit 131 do not necessarily have to be plural, and it is also possible to output allocation data (make all determinations) with one learned model. 〈Shortage calculation unit 132〉

[0070] The shortage calculation unit 132 is a member that outputs data related to beds or medical staff that are insufficient for the treatment of diseases and injuries based on the first input data. Thereby, for example, based on this data, the number of insufficient beds or medical staff can be displayed on the user terminal 12 and transmitted to the user US, and the user US can easily grasp the shortage of beds or medical staff based on this display.

[0071] In addition, the shortage calculation unit 132 can collect data related to beds or medical staff that are insufficient for the treatment of injuries and illnesses in multiple hospitals via the bed control system 1, and generate and output organized data (corresponding to an example of "data related to beds or medical staff that are insufficient for the treatment of injuries and illnesses" in the claims). That is, based on the situations of multiple hospitals, judgments such as "Hospital A lacks beds, but Hospitals B and C have vacant beds, so it is better to transfer the patient." or "Hospital D lacks doctors specializing in a, but Hospital E has doctors specializing in a and no scheduled surgeries today, so it is better to send that doctor for support." can be made. As a result, for example, in the event of an emergency, multiple pieces of information can be instantaneously collected, and requests for patient transfer, doctor support, etc. can be promptly made.

[0072] Note that the mode of the shortage calculation unit 132 is not limited to the above-mentioned one. For example, the method of outputting data of the shortage calculation unit 132 is not particularly limited, and it may use a mathematical model, a learned model, or a combination of these. 〈Input guidance unit 133〉

[0073] The input guidance unit 133 is a member that guides the user US to input information necessary for the first input data as the second input data when the user US uses the user terminal 12 to input data. Note that the second input data is data whose content is equivalent to that of the first input data but whose data format is different from that of the first input data. More specifically, the second input data may include bed data, medical staff data, injury and illness data, patient data, and emergency data, and its formats include text form and voice form.

[0074] When user US inputs the second input data in the form of text using user terminal 12, the input guidance unit 133 communicates with user US in a dialogue form (chat form) using natural language, for example, as shown in FIG. 3, and outputs and displays on user terminal 12 a text for asking for information necessary for the first input data. In FIG. 3, ct1-1 to ct1-6 are displays corresponding to the second input data input by user US, and ct2-1 to ct2-9 are for displaying the text output by the input guidance unit 133.

[0075] For example, in FIG. 3, ct2-1 is a text for asking for the gender of the patient, ct2-3 is a text for asking for the reason for the need for an accompaniment, and ct2-4 is a text for asking for the gender of the accompaniment.

[0076] Note that the mode of the input guidance unit 133 is not limited to the above. For example, if voice synthesis technology is used, it is also possible to replace the above-mentioned chat-form dialogue with a voice dialogue. The same applies when user US inputs the second input data in the form of voice using user terminal 12. Also, the means for determining the information necessary for the first input data and realizing the natural language dialogue are not particularly limited, and those using a mathematical model, a pre-trained model, or a combination of these may be used. 〈Language recognition unit 134〉

[0077] A member composed of a plurality of pre-trained models that are generated by machine learning using third teacher data including a thesaurus of terms and convert the second input data input by user US in the form of text or voice into a form that can be used as the first input data and output it.

[0078] For example, as described in <Input Induction Unit 133>, when second input data is obtained from user US in a dialogue format using natural language, the second input data includes information necessary as first input data and information that is not necessary. In such a case, the language recognition unit 134 extracts only the information necessary as first input data from the second input data, and performs processing to match the extracted data to the format of the first input data. 《Third Teacher Data》

[0079] The third teacher data is data used for generating a pre-trained model that constitutes the language recognition unit 134, and includes a thesaurus of terms (terms are classified and systematized by hierarchical relationships (superordinate and subordinate relationships), part-whole relationships, synonymous relationships, similar relationships, etc.).

[0080] Examples of data that may be included in the third teacher data are given below.

[0081] Thesaurus of general terms: It may include a thesaurus of general terms used in daily conversations.

[0082] Thesaurus of technical terms: It may include a thesaurus of technical terms such as medical terms and nursing terms.

[0083] Thesaurus of hospital-specific terms: It may include a thesaurus of hospital-specific terms (such as jargon and abbreviations). 《Second Shared Pre-trained Model 134a and Second Dedicated Pre-trained Model 134b》

[0084] In addition, the multiple pre-trained models included in the language recognition unit 134 can be classified into a second shared pre-trained model 134a generated by machine learning using data that can be shared among different hospitals in the third teacher data, and a second dedicated pre-trained model 134b generated by machine learning using data that cannot be shared among different hospitals.

[0085] For example, the second shared pre-trained model 134a can be generated using a thesaurus of general terms or a thesaurus of technical terms, and the second dedicated pre-trained model 134b can be generated using a thesaurus of terms unique to the hospital.

[0086] Also, the second shared pre-trained model 134a does not necessarily have to be generated by machine learning using data that can be shared by all hospitals using the bed control system 1. For example, it is understood that this includes cases where it can only be shared among some hospitals, such as between hospitals with similar hospital scales (number of beds), between hospitals with similar hospital characteristics (strong in acute care, cancer specialty, cardiovascular specialty, organ transplant specialty, etc.), between hospitals with similar regional characteristics (regions with many lifestyle diseases due to high sugar and salt intake exceeding the national average, etc.). In this case, it is reasonable because data can be shared between similar hospitals, and a higher learning effect can be expected compared to performing machine learning with the data of a single hospital.

[0087] In this way, by providing the second shared pre-trained model 134a and the second dedicated pre-trained model 134b independently, the second shared pre-trained model 134a can be shared among different hospitals, so that the cost of generating the pre-trained model can be suppressed. In addition, by applying the second dedicated pre-trained model 134b for each hospital, a bed control system capable of handling terms unique to the hospital (jargon, abbreviations, etc.) can be realized. 《Other Aspects of the Language Recognition Unit 134》

[0088] Note that the aspect of the language recognition unit 134 is not limited to the above-mentioned ones.

[0089] For example, the conversion of the second input data into an aspect that can be used as the first input data does not have to be solely borne by the pre-trained model constituting the language recognition unit 134, and a part of it may be shared with other methods such as a mathematical model.

[0090] In addition, it is not necessary to store all of the learned models for the conversion in server 13, and part or all of them may be stored in user terminal 12. For example, second shared learned model 134a may be stored in server 13, and second dedicated learned model 134b may be stored in user terminal 12. That is, the determination that can be shared among different hospitals is made by server 13, and the parts that require unique determination for each hospital are made by user terminal 12 of each hospital, so that the cost of operating the system can be reduced.

[0091] In addition, the learned models that make up bed allocation unit 131 do not necessarily have to be plural, and it is also possible to output allocation data (perform all determinations) using one learned model.

[0092] Furthermore, if it is possible to obtain first input data (or data in a form close thereto) from the beginning, for example, by using a dedicated input form (corresponding to an example of the "input guidance unit" in the claims), language recognition unit 134 does not necessarily have to be provided. 〈Correction unit 135〉

[0093] Correction unit 135 is a member that learns the habits of user US when selecting one bed from among multiple beds when there are multiple beds suitable for allocation to a patient or an attendant, and performs correction based on the habits during subsequent bed control.

[0094] For example, when there are three beds suitable for allocation to a patient or an attendant, bed control system 1 lists the identification information of the three beds together with numbers, as shown in ct2-6 to ct2-8 in FIG. 3. At this time, the higher the degree of suitability (hereinafter referred to as "suitability degree") for allocation to a patient or an attendant, the smaller the number. In other words, bed allocation unit 131 determines that the bed marked with "1." is the most suitable bed.

[0095] On the other hand, as shown at ct1-6 in FIG. 3, user US selects the bed that he actually wants to assign to the patient or the attendant, and the correction unit 135 learns the data related to this selected bed, the first input data, and the fourth teacher data including the assignment data.

[0096] That is, even if there are multiple beds suitable for assignment by the patient or the attendant, if the bed assignment unit 131 can make exactly the same judgment as user US, user US should always select the bed co-written as "1." by the bed assignment unit 131, but this is not actually the case. The correction unit 135 learns the difference between the judgment of user US and the judgment of the bed assignment unit 131, and outputs correction data, which is data for correcting this difference.

[0097] Then, in subsequent bed control, by using this correction data, a plurality of beds can be displayed in a manner that reflects the habits of user US.

[0098] Note that the mode of the correction unit 135 is not limited to the above. Also, the bed control system 1 does not necessarily have to include the correction unit 135. 〈Learning Unit 136〉

[0099] The learning unit 136 is a member that generates a learned model provided in the server 13.

[0100] Note that the mode of the learning unit 136 and the machine learning algorithm are not particularly limited. By periodically regenerating (relearning) the learned model, the accuracy of the learned model can be improved.

[0101] Also, the learning unit 136 does not necessarily have to be provided only in the server 13. It may be provided in the user terminal 12, or may not be provided in the bed control system 1 (in this case, the learned model is generated outside the system). [Advantages of the Present Invention]

[0102] As described above, the bed control system 1 according to an embodiment of the present invention can obtain allocation data in a manner conforming to past bed control by using, for example, a learned model generated by machine learning of past bed control records. That is, when the user US is presented with one bed based on the allocation data, the user only needs to perform a confirmation operation such as whether that bed is acceptable, and when a plurality of beds are presented, the user only needs to perform a selection operation such as selecting one bed from among them, so that bed control becomes possible, and it can be performed more efficiently and with a lower burden.

[0103] Also, unlike the case where bed control automation is realized by a mathematical model or the like, bed control based on reasons that are difficult to verbalize or rule-based can also be learned and reproduced.

Explanation of Signs

[0104] 1…Bed control system 11…Network 12…User terminal 13…Server 131…Bed allocation unit; 131a…First shared learned model; 131b…First dedicated learned model; 131H…Normal-time learned model; 131U…Emergency-time learned model 132…Shortage calculation unit 133…Input guidance unit 134…Language recognition unit; 134a…Second shared learned model; 134b…Second dedicated learned model 135…Correction unit 136…Learning unit US…User AD…Administrator

Claims

1. A bed control system that assigns beds in a hospital to patients or their attendants, comprising a bed allocation unit consisting of one or more trained models that output allocation data, which is data related to one or more beds suitable for allocation to the patient or the attendant, based on first input data including bed data, which is data related to bed information, healthcare worker data, which is data related to healthcare workers' information, injury / sickness data, which is data related to injuries and sicknesses, and patient data, which is data related to patient information, and which is generated by machine learning using first teacher data including the bed data, the healthcare worker data, the injury / sickness data, and the patient data, and second teacher data including data related to one or more beds suitable for allocation to the patient or the attendant based on the first teacher data. A bed control system comprising the same.

2. The bed control system according to Claim 1, wherein the bed allocation unit comprises a first shared trained model generated by machine learning using data that can be shared among different hospitals among the first teacher data and the second teacher data, and a first dedicated trained model generated by machine learning using data that cannot be shared among different hospitals among the first teacher data and the second teacher data. A bed control system comprising the same.

3. The bed control system according to Claim 1, wherein the bed allocation unit comprises an emergency trained model, which is a trained model including emergency data, which is data related to whether there is an emergency in the first teacher data and the first input data, and a normal trained model, which is a trained model not including emergency data in the first teacher data and the first input data. A bed control system comprising the same.

4. The bed control system according to Claim 1, further comprising a shortage calculation unit that outputs data related to beds or healthcare workers that are insufficient for the treatment of injuries and sicknesses based on the first input data.

5. The bed control system according to Claim 1, further comprising a second input unit for the user to input second input data including the patient data in the form of text or voice, and an input guidance unit for guiding the user to input the necessary information for the first input data as the second input data. A language recognition unit comprising one or more trained models that are generated by machine learning using third teacher data including a thesaurus of terms and that convert and output the second input data in a manner that can be used as the first input data. A bed control system comprising the same. **Claim 6** The bed control system according to claim 5, wherein the language recognition unit includes a second shared trained model generated by machine learning using data that can be shared among different hospitals among the third teacher data, and a second dedicated trained model generated by machine learning using data that cannot be shared among different hospitals among the third teacher data. A bed control system comprising the same. **Claim 7** The bed control system according to claim 1, further comprising a display unit that displays a list of a plurality of beds when the bed allocation unit outputs allocation data, which is data related to the plurality of beds suitable for allocation to the patient or the attendant; a selection unit that allows a user to select one bed from among the plurality of beds displayed on the display unit; and a correction unit comprising one or more trained models that are generated by machine learning using fourth teacher data including data related to the plurality of beds and data related to the bed selected by the selection unit, and that output correction data, which is data for correcting the allocation data based on the user's habits. The bed control system comprises the display unit that displays a list of the plurality of beds reflecting the user's habits based on the correction data.

Images