JP7751462B2 - Pharmaceutical inventory management device - Google Patents

Description

The present invention relates to a drug inventory management device.

In hospitals, pharmaceutical inventory management is important because a shortage of pharmaceutical stock can disrupt medical procedures. However, it is extremely time-consuming for hospital staff to manually check inventory levels and record any pharmaceutical inventory reductions. Therefore, a system is known that automatically manages pharmaceutical inventory by attaching a storage medium for storing information, such as an RFID (Radio Frequency Identification) tag, to pharmaceuticals and then reading the information stored on the storage medium when pharmaceutical inventory decreases (see, for example, Patent Document 1).

Japanese Patent Application Laid-Open No. 2021-15577

When managing pharmaceutical inventory, it is important not only to obtain information on pharmaceutical inventory levels, but also to accurately predict how much inventory should be maintained when replenishing in the first place, in order to reduce the risk of stockouts while also reducing excess inventory and cutting costs. However, while the invention described in Patent Document 1 was able to automatically record decreases in inventory levels, it was unable to predict how much inventory should be maintained.

An object of the present invention is to provide a drug inventory management device that can predict the number of drugs to be stocked in a hospital.

In order to solve the above problem, the invention described in claim 1 is a pharmaceutical inventory management device ,
A prediction model generation/acquisition means for generating or acquiring a prediction model that predicts the number of pharmaceuticals used in a hospital;
a usage quantity prediction means for predicting the usage quantity of a target drug, which is a drug that is the subject of prediction, at a target hospital, which is a hospital that is the subject of prediction, during a target period, which is a period that is the subject of prediction, using the prediction model;
an appropriate inventory quantity prediction means for predicting the inventory quantity of the target drug that should be secured in the target hospital by adding a preset value to the usage quantity of the target drug predicted by the usage quantity prediction means;
Equipped with
The prediction model generation/acquisition means generates the prediction model using information linking at least information related to the type of drug and information related to the characteristics of the hospital with information related to the number of drugs used in a predetermined period, or acquires the prediction model generated using information linking at least information related to the type of drug and information related to the characteristics of the hospital with information related to the number of drugs used in a predetermined period,
The usage quantity prediction means inputs at least information related to the type of the target drug, information related to the characteristics of the target hospital , information related to the target period , and information related to the characteristics of patients hospitalized at the target hospital into the prediction model, and predicts the usage quantity of the target drug at the target hospital during the target period.

In order to solve the above problem, the invention described in claim 2 is a pharmaceutical inventory management device ,
A prediction model generation/acquisition means for generating or acquiring a prediction model that predicts the number of pharmaceuticals used in a hospital;
a usage quantity prediction means for predicting the usage quantity of a target drug, which is a drug that is the subject of prediction, at a target hospital, which is a hospital that is the subject of prediction, during a target period, which is a period that is the subject of prediction, using the prediction model;
an appropriate inventory quantity prediction means for predicting the inventory quantity of the target drug that should be secured in the target hospital by adding a preset value to the usage quantity of the target drug predicted by the usage quantity prediction means;
Equipped with
The prediction model generation/acquisition means generates the prediction model using information linking at least information related to the type of drug and information related to the number of drugs used in the target hospital for a predetermined period of time, or acquires the prediction model generated using information linking at least information related to the type of drug and information related to the number of drugs used in the target hospital for a predetermined period of time,
The usage quantity prediction means is characterized by inputting at least information regarding the type of the target drug , information regarding the target period , and information regarding the characteristics of patients hospitalized at the target hospital into the prediction model, and predicting the number of times the target drug will be used at the target hospital during the target period.

The invention described in claim 3 is the pharmaceutical inventory management device described in claim 1 or 2 ,
The usage quantity prediction means further inputs information relating to the season to which the target period belongs into the prediction model, and predicts the usage quantity of the target drug at the target hospital during the target period.

The invention described in claim 4 is the pharmaceutical inventory management device described in any one of claims 1 to 3 ,
The usage quantity prediction means further inputs information related to the location of the target hospital into the prediction model, and predicts the usage quantity of the target drug at the target hospital during the target period.

The invention described in claim 5 is the pharmaceutical inventory management device according to any one of claims 1 to 4 ,
The predictive model generation/acquisition means generates the predictive model by machine learning using data linked at least to information related to the type of pharmaceutical product and information related to the number of pharmaceutical products used as training data, or acquires the predictive model generated by machine learning using data linked at least to information related to the type of pharmaceutical product and information related to the number of pharmaceutical products used as training data.

The invention described in claim 6 is the pharmaceutical inventory management device according to any one of claims 1 to 5 ,
The value added by the optimum inventory quantity prediction means is a value set for each type of pharmaceutical product, and is characterized in that the value added is set to be larger for pharmaceutical products that are used frequently, pharmaceutical products with long expiration dates, and pharmaceutical products whose shortage may pose a risk to the patient's life than for other pharmaceutical products .

The invention described in claim 7 is the pharmaceutical inventory management device according to any one of claims 1 to 6 ,
The value added by the optimum stock quantity prediction means is a value set by the target hospital.

According to the present invention, it is possible to provide a drug inventory management device that can predict the number of drugs to be stocked in a hospital.

1 is a diagram illustrating an example of a pharmaceutical shop in which an article management system according to an embodiment is installed. 1 is a block diagram showing the configuration of an article management system according to an embodiment; 1 is a flowchart showing an outline of the flow of operations of an item management system according to an embodiment. 10 is a flowchart showing details of the operational flow of entry processing, take-out processing, and exit processing of the article management system according to the embodiment. 10 is a flowchart showing details of the operation flow of the optimum inventory quantity prediction process of the article management system according to the embodiment. FIG. 1 is a diagram showing the flow of data when generating a prediction model and predicting usage quantities in an item management system according to an embodiment.

The following describes an item management system 100, an embodiment of the present invention, with reference to Figures 1 to 6. However, the technical scope of the present invention is not limited to the illustrated example.

[First Configuration]
The article management system 100 is a system for managing the inventory of medicines stored in a medicine shop S established within a hospital, and as shown in FIGS. 1 and 2, comprises an article management device 1 and a take-out processing RFID reader 3 installed within the medicine shop S, an entry processing RFID reader 2 and an exit processing RFID reader 4 installed at an entrance/exit E of the medicine shop S, and staff RFID tags (staff RFID tags 5a to 5c, hereinafter, any staff RFID tag will be referred to as staff P when no particular distinction is made) carried by each staff member of the hospital in which the medicine shop S is established (staff members P1 to P3, hereinafter, any staff member will be referred to as staff P when no particular distinction is made). The inventory management system 100 is configured to include: pharmaceutical RFID tags (pharmaceutical RFID tags 6a to 6c; hereinafter, when referring to any pharmaceutical RFID tag without any distinction, it will be referred to as pharmaceutical RFID tag 6) attached to each pharmaceutical (pharmaceuticals M1 to M3; hereinafter, when referring to any pharmaceutical without any distinction, it will be referred to as pharmaceutical RFID tag 6) in the pharmaceutical shop S; a wholesaler server 7 which is a server outside the pharmaceutical shop S used by a wholesaler that stocks pharmaceuticals in the pharmaceutical shop S; and a central server 8 which is a server outside the pharmaceutical shop S used by the administrator who manages the item management system 100.
In addition, the item management device 1 is connected to the RFID reader 2 for entry processing, the RFID reader 3 for removal processing, and the RFID reader 4 for exit processing via a communication cable C, and the item management device 1 is connected to the wholesaler server 7 and the wholesaler server 7 and the central server 8 via a communication network N.

In Figure 2, for ease of explanation, only three staff members, P1 to P3, and three medicines, M1 to M3, are shown, but in actual operation, many more medicines will be placed in the medicine shop S and used by many more staff members.
2 shows only one item management device 1 and one wholesaler server 7, an item management device 1 is installed in each of multiple hospitals that use this system and have a pharmaceutical shop S. In addition, if there are multiple wholesalers that wholesale pharmaceuticals M, such as when different wholesalers are used for each hospital, a wholesaler server 7 will be installed individually for each of the multiple wholesalers.

The pharmaceutical shop S is a room in the hospital where a specific wholesaler who distributes pharmaceuticals M to the hospital keeps pharmaceuticals M. Note that, although Fig. 1 illustrates a case where there is no passage connecting the inside and outside of the pharmaceutical shop S other than the entrance/exit E that serves as both the entrance and exit to the pharmaceutical shop S, a separate entrance and exit may be provided.
Furthermore, pharmaceutical products M include medicines placed in pharmaceutical shop S by pharmaceutical wholesalers, as well as medical equipment such as hemostatic products, surgical tools, and syringes, and are not limited to drugs as long as they are used in hospitals.
In addition, hospital staff P broadly includes doctors, nurses, other staff, and others who work at the hospital where the pharmaceutical shop S is located.

While ownership of pharmaceuticals M placed within pharmaceutical shop S depends on the contract between the wholesaler and the hospital where pharmaceutical shop S is located, it is preferable that the pharmaceuticals M be retained by the wholesaler while placed within pharmaceutical shop S, and transferred to the hospital when they are taken out of pharmaceutical shop S through entrance/exit E by staff P, as described below.

[1 Article management device]
The item management device 1 is, for example, an information device such as a PC (Personal Computer) or WS (Work Station) installed in a pharmaceutical shop S, and as described below, acquires and stores information read from an entry processing RFID reader 2, an exit processing RFID reader 3, and an exit processing RFID reader 4, using these to read from an RFID tag 5 for staff and an RFID tag 6 for pharmaceuticals.
In addition, the goods management device 1 communicates with the wholesaler server 7 via the communication network N to send and receive information as described below.
As shown in FIG. 2, the article management device 1 includes, for example, a control unit 11, a storage unit 12, a communication unit 13, a display unit 14, and an operation unit 15.

[(1) Control Unit]
The control unit 11 is a part that controls the operation of the item management device 1, and is configured with, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc., and controls each part of the item management device 1 in cooperation with the CPU and the program data stored in the memory unit 12.

[(2) Storage section]
The memory unit 12 is a part where various information necessary for the operation of the item management device 1 is stored, and is composed of, for example, an HDD (Hard Disk Drive), semiconductor memory, etc., and stores data necessary for the operation of the item management device 1 in a manner that allows it to be read and written by the control unit 11.
The memory unit 12 stores a program including various commands to the control unit 11 for operating the item management device 1, and the operation of the item management device 1 described in the operation explanation below is performed in accordance with the program stored in the memory unit 12.

The memory unit 12 also includes a staff database 121 that stores information related to staff P, a pharmaceutical database 122 that stores information related to pharmaceuticals M, and a staff member entering/pharmaceutical member taking out information memory unit 123 that stores information related to staff P who enter the pharmaceutical shop S and pharmaceuticals M taken out by staff P, as described below.

The staff database 121 stores information related to each staff member's name, affiliation (hospital, department, medical department), etc., linked to staff identification information (staff identification information D1a to D1c; hereinafter, when referring to any of the staff identification information without making any distinction, it will be referred to as staff identification information D1), which is a unique identification number assigned to each staff member P stored on the staff RFID tag 5 carried by each staff member P.
In addition, the pharmaceutical database 122 stores information related to the name, product category, manufacturer name, sales price, name of the wholesaler that delivered the pharmaceutical, etc., linked to pharmaceutical identification information (pharmaceutical identification information D2a to D2c; hereinafter, when referring to any of the pharmaceutical identification information without making any particular distinction, it will be referred to as pharmaceutical identification information D2), which is a unique identification number assigned to each pharmaceutical M stored in the pharmaceutical RFID tag 6 attached to each pharmaceutical M.
The pharmaceutical database 122 also stores information on the number of pharmaceuticals in stock at the pharmaceutical shop S for each type (name) of pharmaceutical.

[(3) Communications Department]
The communication unit 13 is a part used for communication between the goods management device 1 and the RFID reader 2 for entry processing, the RFID reader 3 for removal processing, the RFID reader 4 for exit processing, and the wholesaler server 7, and is a communication interface having, for example, a communication IC (Integrated Circuit) and a communication connector, and under the control of the control unit 11, performs data communication between the RFID reader 2 for entry processing, the RFID reader 3 for removal processing, and the RFID reader 4 for exit processing via the communication cable C, and data communication with the wholesaler server 7 via the communication network N using a predetermined communication protocol.

[(4) Display section]
The display unit 14 includes a display such as an LCD (Liquid Crystal Display), and displays an image based on a display control signal output from the control unit 11 on the display screen.

[(5) Operation section]
The operation unit 15 includes, for example, a keyboard having character input keys, number input keys, and other keys associated with various functions, and outputs operation signals according to operation inputs using these keys to the control unit 11. The operation unit 15 may be, for example, a touch panel formed integrally with the display unit 14.

[2. RFID reader for entry processing]
The entry processing RFID reader 2 is a reading device installed at the entrance of the pharmaceutical shop S (in this embodiment, the entrance/exit E, which serves as both an entrance and an exit), and is used to read the staff identification information D1 stored in the staff RFID tag 5 when staff P enters the pharmaceutical shop S, as described below.

The entry processing RFID reader 2 comprises a reader body 21 and a sheet antenna 22 that communicates with the staff RFID tag 5.

The reader main body 21 reads information stored in the staff RFID tag 5 from the radio waves received by the sheet antenna 22 from the tag.
In addition, the reader main body 21 is equipped with a control unit 211 that controls the operation of the reader main body 21 and a communication unit 212 that communicates with the item management device 1, and is connected to the item management device 1 via a communication cable C.

The sheet antenna 22 is, for example, a rectangular sheet-shaped antenna, and as described below, receives radio waves emitted from an employee RFID tag 5 held over the sheet antenna 22.

[3. RFID reader for removal processing]
The RFID reader 3 for take-out processing is a reading device installed in the pharmaceutical shop S, and is used to read the staff identification information D1 stored in the staff RFID tag 5 and the pharmaceutical identification information D2 stored in the pharmaceutical RFID tag 6 when the staff P takes out the pharmaceutical M, as described below.

The RFID reader 3 for takeout processing comprises a reader body 31 and a sheet antenna 32 that communicates with the staff RFID tag 5 and the medicine RFID tag 6.

The reader main body 31 reads the information stored in the staff RFID tag 5 and the medicine RFID tag 6 from the radio waves received by the sheet antenna 32 from these tags.
In addition, the reader main body 31 is equipped with a control unit 311 that controls the operation of the reader main body 31 and a communication unit 312 that communicates with the item management device 1, and is connected to the item management device 1 via a communication cable C.

The sheet antenna 32 is, for example, a rectangular sheet-shaped antenna with an area large enough to place a pharmaceutical product M on it, and as described below, receives radio waves emitted from a staff RFID tag 5 held over the sheet antenna 32 and from a pharmaceutical product RFID tag 6 attached to a pharmaceutical product M placed on the sheet antenna 32.

[4. RFID reader for exit processing]
The exit processing RFID reader 4 is a reading device installed at the exit of the pharmaceutical shop S (in this embodiment, the entrance/exit E, which serves as both the entrance and exit), and is used to read the staff identification information D1 stored in the staff RFID tag 5 and the pharmaceutical identification information D2 stored in the pharmaceutical RFID tag 6 when the staff member P leaves the pharmaceutical shop S, as described below.

The exit processing RFID reader 4 comprises a reader body 41 and a sheet antenna 42 that communicates with the staff RFID tag 5 and the medicine RFID tag 6.

The reader main body 41 reads the information stored in the staff RFID tag 5 and the medicine RFID tag 6 from the radio waves received by the sheet antenna 42 from these tags.
In addition, the reader main body 41 is equipped with a control unit 411 that controls the operation of the reader main body 41 and a communication unit 412 that communicates with the item management device 1, and is connected to the item management device 1 via a communication cable C.

The sheet antenna 42 is, for example, a rectangular sheet-shaped antenna, and as described below, receives radio waves emitted from the staff RFID tag 5 carried by staff member P as he leaves the pharmaceutical shop S through the entrance/exit E, and from the pharmaceutical RFID tag 6 attached to the pharmaceutical M that the staff member takes out.

In addition, in the case where the entrance and exit to the pharmaceutical shop S are common (entrance/exit E), as in the pharmaceutical shop S of this embodiment, a single RFID reader may serve as both the RFID reader 2 for entry processing and the RFID reader 4 for exit processing.
In this case, such a single RFID reader serves as both the second reading device and the third reading device of the present invention.
In contrast to this, unlike this embodiment, if the pharmaceutical shop S has a separate entrance for staff P to enter and an exit for staff P to exit, an RFID reader 2 for entry processing will be provided at the entrance, and an RFID reader 4 for exit processing will be provided at the exit.

[5. Staff RFID tags]
The staff RFID tag 5 is an RFID tag carried by each staff member P, and stores staff identification information D1, which is a unique identification number assigned to each staff member P carrying the staff RFID tag 5.

The staff RFID tag 5 is, for example, a passive RFID tag without a built-in battery, and as shown in Figure 2, includes an antenna unit 51 (51a to 51c) and an IC chip 52 (52a to 52c).

The antenna unit 51 receives radio waves emitted from the sheet antenna 22 of the RFID reader 2 for entry processing, the sheet antenna 32 of the RFID reader 3 for removal processing, or the sheet antenna 42 of the RFID reader 4 for exit processing, converts the received radio waves into electricity, and supplies power to the IC chip 52.
The IC chip 52 has an internal memory that stores staff identification information D1, and when the antenna unit 51 receives radio waves emitted from the sheet antenna 22 of the RFID reader 2 for entry processing, the sheet antenna 32 of the RFID reader 3 for removal processing, or the sheet antenna 42 of the RFID reader 4 for exit processing, the IC chip 52 controls the antenna unit 51 to radially emit radio waves including the staff identification information D1.

[6. RFID tags for pharmaceuticals]
The medicine RFID tag 6 is an RFID tag provided on each medicine M, and stores medicine identification information D2, which is a unique identification number given to each medicine M on which the medicine RFID tag 6 is provided.

The pharmaceutical RFID tag 6 is, for example, a passive RFID tag without a built-in battery, and as shown in Figure 2, includes an antenna unit 61 (61a to 61c) and an IC chip 62 (62a to 62c).

The antenna unit 61 receives radio waves transmitted from the sheet antenna 32 of the RFID reader 3 for take-out processing or the sheet antenna 42 of the RFID reader 4 for exit processing, converts the received radio waves into electricity, and supplies the electricity to the IC chip 62 .
The IC chip 62 has an internal memory that stores pharmaceutical identification information D2, and when the antenna unit 61 receives radio waves emitted from the sheet antenna 32 of the RFID reader 3 for take-out processing or the sheet antenna 42 of the RFID reader 4 for exit processing, the IC chip 62 controls the antenna unit 61 to radially emit radio waves including the pharmaceutical identification information D2.

[7 Wholesaler Server]
The wholesaler server 7 is, for example, an information device such as a PC or WS owned by a wholesaler that stocks pharmaceuticals in the pharmaceutical shop S, and as described below, transmits information related to the results of tallying the number of times pharmaceuticals M are used to the central server 8.
As shown in FIG. 2, the wholesaler server 7 includes, for example, a control unit 71, a storage unit 72, and a communication unit 73.

The configurations of the control unit 71 and communication unit 73 are identical to the control unit 11 and communication unit 13, respectively, in the item management device 1.

The memory unit 72 is configured, for example, with an HDD, semiconductor memory, etc., similar to the memory unit 12 in the goods management device 1, and stores data required for the operation of the wholesaler server 7 in a manner that allows the control unit 71 to read and write the data.
The memory unit 72 stores a program including various commands to the control unit 71 for operating the wholesaler server 7, and the operation of the wholesaler server 7 described in the operation explanation below will be performed in accordance with the program stored in the memory unit 72.

[8. Central Server]
The central server 8 is, for example, an information device such as a PC or WS owned by an administrator who manages the entire inventory management system 100, and performs processes related to generating a prediction model 823, predicting the number of pharmaceuticals to be used using the prediction model 823, and predicting the number of pharmaceuticals to be stocked at the hospital by placing them in the pharmaceutical shop S, as described below.
As shown in FIG. 2, the central server 8 includes, for example, a control unit 81, a storage unit 82, and a communication unit 83.

The configurations of the control unit 81 and communication unit 83 are identical to those of the control unit 11 and communication unit 13 in the item management device 1, respectively.

The memory unit 82, like the memory unit 12 in the item management device 1, is composed of, for example, an HDD, semiconductor memory, etc., and stores data required for the operation of the central server 8 in a manner that allows it to be read and written by the control unit 81.

The storage unit 82 includes a count result information storage unit 821 in which the count result information D7 is stored, and an appropriate stock quantity information storage unit 824 in which the appropriate stock quantity information D11 is stored.
In addition, the memory unit 82 stores a prediction model generation program 822, which is a program including a machine learning algorithm used when generating the prediction model 823, as described below, and after the prediction model 823 is generated, the prediction model 823 is also stored.

In addition to the above, the memory unit 82 stores a program containing various commands to the control unit 81 for operating the central server 8, and the operation of the central server 8 described below in the operation explanation is performed in accordance with the program stored in the memory unit 82.

[9 Communication Cable]
The communication cable C is, for example, a cable such as a USB (Universal Serial Bus) cable or a coaxial cable, and connects the item management device 1 with the entry processing RFID reader 2, the take-out processing RFID reader 3, and the exit processing RFID reader 4, as shown in Figures 1 and 2.

The communication cable C is not particularly limited as long as it can connect the devices that make up the item management system 100 as described above and can transmit and receive data between them.
Furthermore, instead of the communication cable C, these devices may be connected by wireless communication.

[10 Communication Network]
The communication network N is, for example, the Internet, a telephone line network, a mobile phone communication network, a wireless LAN communication network, etc., and connects the goods management device 1 and the wholesaler server 7, and between the wholesaler server 7 and the central server 8, as shown in Figure 2.
The communication network N is not particularly limited as long as it can connect the item management device 1 and the wholesaler server 7, and between the wholesaler server 7 and the central server 8, as described above, and can transmit and receive data between them.

[Second Description of Operation]
The operation of the item management system 100 according to this embodiment will be described below. As shown in the flowchart of Figure 3, the operation of the item management system 100 is roughly divided into five steps: entry processing (step S1), removal processing (step S2), exit processing (step S3), usage count processing (step S4), and optimal inventory quantity prediction processing (step S5).

[1 Step S1: Entry Process]
First, the details of the operation of this system when staff member P enters the medicine shop S will be described with reference to the flowchart of FIG.

When a staff member P enters the drug shop S through an entrance E, he or she holds up the staff RFID tag 5 he or she carries over the sheet antenna 22 of the entrance processing RFID reader 2 provided at the entrance E.
When the staff RFID tag 5 is held over the sheet antenna 22, in the entry processing RFID reader 2, the sheet antenna 22 receives the radio waves emitted from the staff RFID tag 5, and the control unit 211 of the reader main body 21 reads the staff identification information D1 stored in the staff RFID tag 5 held over the sheet antenna 22 from the radio waves received by the sheet antenna 22 (step S1-1).

After reading the staff identification information D1, the control unit 211 of the reader main body 21 of the entry processing RFID reader 2 transmits the read staff identification information D1 from the communication unit 212 to the item management device 1 via the communication cable C (step S1-2).

In the item management device 1, which receives the staff member identification information D1 transmitted from the reader main body 21 of the entry processing RFID reader 2 via the communication unit 13, the control unit 11 stores the received staff member identification information D1 in the entering staff member/takeout medicine information storage unit 123 of the storage unit 12 (step S1-3). At this time, the control unit 11 also associates information regarding the time when the staff member identification information D1 was received in step S1-2 (entry time information D4) with the staff member identification information D1 and stores it in the entering staff member/takeout medicine information storage unit 123 of the storage unit 12.

As a result, each time staff member P enters the pharmaceutical shop S, staff member identification information D1 related to that staff member P is stored in the memory unit 12 of the item management device 1.

[2. Step S2: Take-out Processing]
Next, after the staff member P selects the medicine M that he/she wishes to take out, the system performs the take-out processing using the take-out processing RFID reader 3, i.e., the process of having the take-out processing RFID reader 3 read the staff RFID tag 5 carried by the staff member P and the medicine RFID tag 6 attached to the medicine M to be taken out. The details of the operation of this system will be explained using the flowchart in Figure 4.

When a staff member P selects a medicine M that he/she wishes to take out from among the medicines M placed in the medicine shop S, the staff member P takes the medicine M to the RFID reader 3 for take-out processing and holds up the staff RFID tag 5 that he/she carries over the sheet antenna 32 of the RFID reader 3 for take-out processing.
When the staff RFID tag 5 is held over the sheet antenna 32, in the RFID reader 3 for takeout processing, the sheet antenna 32 receives the radio waves emitted from the staff RFID tag 5, and the control unit 311 of the reader main body 31 reads the staff identification information D1 stored in the staff RFID tag 5 held over the sheet antenna 32 from the radio waves received by the sheet antenna 32 (step S2-1).

Next, the staff member P places the medicine M that he/she wishes to take out on the sheet antenna 32 of the RFID reader 3 for take-out processing.
When the pharmaceutical product M is placed on the sheet antenna 32, in the RFID reader 3 for take-out processing, the sheet antenna 32 receives radio waves emitted from the pharmaceutical product RFID tag 6 attached to the pharmaceutical product M, and the control unit 311 of the reader main body 31 reads the pharmaceutical product identification information D2 stored in the pharmaceutical product RFID tag 6 attached to the pharmaceutical product M placed on the sheet antenna 32 from the radio waves received by the sheet antenna 32 (step S2-2).

The order of steps S2-1 and S2-2 may be reversed. Alternatively, the staff member P may hold both the staff RFID tag 5 and the drug RFID tag 6 over the sheet antenna 32 at the same time, and the control unit 311 of the reader main body 31 may read both the staff member identification information D1 and the drug identification information D2 at the same time.
Alternatively, the staff member P1 may hold multiple medicines M over the sheet antenna 32 simultaneously or sequentially, and the control unit 311 of the reader main body 31 may read multiple pieces of medicine identification information D2 simultaneously or sequentially.

After reading the staff identification information D1 and drug identification information D2, the control unit 311 of the reader body 31 of the RFID reader 3 for takeout processing links the read staff identification information D1 and drug identification information D2 and transmits them from the communication unit 312 to the item management device 1 via the communication cable C (step S2-3).

In the item management device 1, which receives the staff identification information D1 and pharmaceutical identification information D2 transmitted from the reader main body 31 of the RFID reader 3 for take-out processing via the communication unit 13, the control unit 11 stores the received pharmaceutical identification information D2 in the entering staff/take-out pharmaceutical information storage unit 123 of the storage unit 12, linking it to the staff identification information D1 stored in step S1-3 that matches the staff identification information D1 received in step S2-3 (step S2-4).
At this time, the control unit 11 also links information regarding the time at which the staff identification information D1 and pharmaceutical identification information D2 were received in step S2-3 (take-out processing time information D5) to the staff identification information D1 and pharmaceutical identification information D2, and stores it in the staff entering/take-out pharmaceutical information memory unit 123 of the memory unit 12.

As a result, each time staff member P performs a take-out process for a medicine M, the medicine identification information D2 for the medicine M that was taken out is linked to the staff member identification information D1 for the staff member P who takes out the medicine M, and is stored in the memory unit 12 of the item management device 1.

[3 Step S3: Exit Processing]
Next, the operation of the system when the staff member P leaves the drug shop S with the drug M in hand will be described in detail with reference to the flowchart of FIG.

When staff member P passes through entrance/exit E carrying pharmaceuticals M, the sheet antenna 42 of the exit processing RFID reader 4 installed at entrance/exit E receives radio waves emitted from the staff RFID tag 5 and the pharmaceutical RFID tag 6, and the control unit 411 of the reader main body 41 reads, from the radio waves received by the sheet antenna 42, staff member identification information D1 stored in the staff RFID tag 5 carried by staff member P passing through entrance/exit E and pharmaceutical identification information D2 stored in the pharmaceutical RFID tag 6 attached to the pharmaceutical M taken out by said staff member P (step S3-1).

After reading the staff identification information D1 and drug identification information D2, the control unit 411 of the reader body 41 of the room exit processing RFID reader 4 links the read staff identification information D1 and drug identification information D2 and transmits them from the communication unit 412 to the item management device 1 via the communication cable C (step S3-2).

In the item management device 1, which receives the staff identification information D1 and pharmaceutical identification information D2 transmitted from the reader main body 41 of the RFID reader 4 for exit processing via the communication unit 13, the control unit 11 determines whether information matching the pharmaceutical identification information D2 received in step S3-2 is stored in the staff entering/pharmaceutical information storage unit 123 of the storage unit 12 (step S3-3).
That is, the control unit 11 determines whether the pharmaceutical identification information D2 received in step S3-2 matches the pharmaceutical identification information D2 received from the take-out processing RFID reader 3 in step S2-3 and stored in the entering staff/take-out pharmaceutical information storage unit 123 of the storage unit 12 in step S2-4, and if such information is found, it further determines whether the staff identification information D1 stored in association with such pharmaceutical identification information D2 matches the staff identification information D1 received in step S3-2, and if there is a match, it determines that information matching the pharmaceutical identification information D2 received in step S3-2 is stored in the entering staff/take-out pharmaceutical information storage unit 123 of the storage unit 12.

If it is determined in step S3-3 that information matching the pharmaceutical identification information D2 received in step S3-2 is stored in the staff member entering/drug information storage unit 123 of the storage unit 12, the control unit 11 associates information regarding the time at which the staff member identification information D1 and pharmaceutical identification information D2 were received in step S3-2 (exit time information D6) with the staff member identification information D1 and stores it in the staff member entering/drug information storage unit 123 of the storage unit 12 (step S3-4).

In contrast, if it is determined in step S3-3 that information matching the pharmaceutical identification information D2 received in step S3-2 is not stored in the staff entering/pharmaceutical information storage unit 123 of the storage unit 12, the control unit 11 of the item management device 1 links the pharmaceutical identification information D2 and exit time information D6 received in step S3-2 to the staff identification information D1 stored in step S1-3 that matches the information received in step S3-2, and stores it in the staff entering/pharmaceutical information storage unit 123 of the storage unit 12 (step S3-5).
At this time, the control unit 11 also links the pharmaceutical identification information D2 and stores predetermined information (uncompleted take-out flag D3) indicating that the pharmaceutical identification information D2 relates to pharmaceuticals for which the take-out process of step S2 has not been performed in the entering staff/take-out pharmaceutical information storage unit 123 of the storage unit 12.

As a result, when staff member P leaves the pharmaceutical shop S through entrance/exit E with pharmaceutical M in hand, it is determined whether the take-out process of step S2 has been performed for that pharmaceutical M. Even if pharmaceutical M is taken out without the take-out process of step S2 being performed, pharmaceutical identification information D2 for the taken-out pharmaceutical M is linked to staff member identification information D1 for staff member P who took out the pharmaceutical, and stored in the memory unit 12 of the item management device 1. At this time, a take-out process incomplete flag D3 is also stored linked to pharmaceutical identification information D2, making it possible to identify pharmaceutical M taken out without the take-out process of step S2 being performed later.

[4. Step S4: Usage Counting Process]
Next, the operation of the system when tallying up the number of medicines taken out of the medicine shop S and used will be described in detail.

In the item management device 1, the control unit 11 compiles, at predetermined intervals (for example, once a week, once a month, etc.), the medicines M that have been taken out of the pharmaceutical shop S during that period and for which new pharmaceutical identification information D2 has been stored in the staff entering/taken out pharmaceutical information memory unit 123 of the memory unit 12, by type (name) of pharmaceutical.
Specifically, the pharmaceutical database 122 stores information relating to the names of corresponding pharmaceuticals linked to pharmaceutical identification information D2, and from this information, the names of the corresponding pharmaceuticals can be identified for each piece of pharmaceutical identification information D2 stored in the staff entering/drug taking-out information storage unit 123 during a specified period, and then the data can be compiled by pharmaceutical name.

Once the tallying is complete, the control unit 11 of the item management device 1 transmits tally result information D7, which is information linking the tally results for each name of the medicine M taken out of the medicine shop S during a specified period with the identification information of the hospital where the item management device 1 is installed, from the communication unit 13 to the wholesaler server 7 via the communication network N.

In the wholesaler server 7, which receives the tally result information D7 sent from the goods management device 1 via the communication unit 73, the control unit 71 transmits the tally result information D7 from the communication unit 73 to the central server 8 via the communication network N.
In the central server 8, which receives the aggregation result information D7 sent from the wholesaler server 7 via the communication unit 83, the control unit 81 stores the received aggregation result information D7 in the aggregation result information memory unit 821 of the memory unit 82.

As a result, the central server 8 will store information related to the results of tallying the number of pharmaceuticals used (the number of each pharmaceutical M taken out of the pharmaceutical shop S) for each specified period, such as one week or one month, and by type (name) of pharmaceutical, at each hospital where the pharmaceutical shop S is located.

[5. Step S5: Optimal Stock Quantity Prediction Process]
Next, the details of the operation of this system when predicting the inventory quantity of each type (name) of pharmaceutical M that should be secured at the hospital where the pharmaceutical shop S is located by placing it in the pharmaceutical shop S will be explained with reference to Figures 5 and 6.
As shown in FIG. 5, this step consists of three steps: step S5-1 (generation of a prediction model), step S5-2 (prediction of the number of uses), and step S5-3 (prediction of the optimum inventory number).

(1) Step S5-1: Generation of Prediction Model
First, the flow of generating a prediction model 823 for predicting the number of drugs used in the central server 8 will be described with reference to FIG.

It is estimated that there is a certain correlation between information on the type (name) of pharmaceuticals and information on the characteristics of the hospital (e.g., size (number of beds), types of illnesses treated, etc.), and the number of pharmaceuticals used at the hospital. In other words, for example, if the number of uses of a pharmaceutical for a specific illness is tallied at hospitals with a department that treats a large number of those illnesses, the number of uses is expected to be high, and if the number of uses is tallied at hospitals that do not have such a department, the number of uses is expected to be low.

Therefore, in step S5-1, the central server 8 generates a prediction model 823 as an example of a prediction model for predicting the number of pharmaceuticals used in a hospital, which can predict the number of pharmaceuticals used in the hospital within that period when information related to at least the type of pharmaceutical, information related to the characteristics of the hospital, and information related to the period to be predicted is input.

Specifically, the central server 8 first creates training data D8 to be used for machine learning. That is, as described in step S4, the count result information storage unit 821 of the storage unit 82 of the central server 8 stores, as count result information D7, information related to the count results of the number of pharmaceuticals used for each type (name) of pharmaceutical for each predetermined period in each hospital where the pharmaceutical shop S is installed.
Therefore, based on the aggregation result information D7, the control unit 81 creates training data D8, which is data linked to at least information regarding the type (name) of the drug M (drug type information D8-2) and information regarding the characteristics of the hospital (hospital characteristics information D8-3), and information regarding the number of times that type of drug M is used in the hospital per specified period (drug usage number information D8-1).

Hospital characteristic information D8-3 includes information such as the size of the hospital (number of beds), the types of illnesses it can treat, and the types of medical departments it offers. Information about the characteristics of each hospital in which a pharmaceutical shop S is located may be acquired in advance and stored in the memory unit 82, or it may be acquired externally and used one by one.

When the training data D8 is created, the control unit 81 inputs the data into a prediction model generation program 822, which is a program that includes a predetermined machine learning algorithm, as shown in Figure 6, and generates a prediction model 823 by having the data learn.
The machine learning algorithm included in the prediction model generation program 822 is not particularly limited, and a known algorithm such as a neural network can be used, for example.

Note that the information linked to the drug usage quantity information D8-1 in the training data D8 is not limited to drug type information D8-2 and hospital characteristics information D8-3.

For example, in addition to the drug type information D8-2 and hospital characteristic information D8-3, seasonal information (seasonal information D8-4) linked to the drug usage quantity information D8-1 may be used. For example, the seasonal information D8-4 may define March to May as spring, June to August as summer, September to November as autumn, and December to February as winter, and use information regarding which season each period for which the usage quantity of drug M in the drug usage quantity information D8-1 is tallied belongs to (if the period spans two seasons, the longer of the seasons may be used).

Furthermore, for example, in addition to the pharmaceutical type information D8-2 and hospital characteristic information D8-3, information relating to the characteristics of patients who were actually hospitalized in the hospital (patient characteristic information D8-5) may be used as data linked to pharmaceutical usage quantity information D8-1.
As the patient characteristic information D8-5, for example, the central server 8 may obtain and use information regarding the number of patients hospitalized in the hospital by symptom category during each period in which the number of uses of the medicine M was tallied (for example, by dividing into injuries and illnesses, and then further dividing into categories by specific body parts and symptoms, and the number of patients in each category) via the wholesaler server 7 or directly from the goods management device 1.

Furthermore, for example, in addition to the drug type information D8-2 and the hospital characteristic information D8-3, information relating to the location of the hospital (hospital location information D8-6) may be used as data linked to the drug usage quantity information D8-1.
The hospital location information D8-6 may, for example, be information regarding the area in which the hospital where the usage of the drug M has been tallied is located within a subdivided area such as a prefecture or city, or information regarding the area in which the hospital is located within a broad area such as the Kanto region or the Kansai region.

Note that the season information D8-4, patient characteristic information D8-5, and hospital location information D8-6 may all be included in the training data D8, or only one or two of them may be included in the training data D8. Figure 6 illustrates a case where all of these are included in the training data D8.

In the central server 8, the prediction model 823 as described above is generated periodically using the latest information, for example, once every two weeks, once a month, etc. Specifically, for example, all of the aggregation result information D7 stored up to that point in the aggregation result information storage unit 821 of the storage unit 82 may be used, or a certain range of information from the most recent aggregation result information D7 stored in the aggregation result information storage unit 821 may be used.
This allows the prediction model 823 to be always updated based on the latest data.

Once the prediction model 823 is generated, the control unit 81 of the central server 8 stores the generated prediction model 823 in the memory unit 82.

[(2) Step S5-2: Predicting usage]
Next, the flow of the process in which the central server 8 predicts the number of pharmaceuticals to be used in a hospital using the prediction model 823 generated in step S5-1 will be described with reference to FIG.

In this case, first, the administrator of the system sets the type (name) of the drug to be predicted (hereinafter referred to as the "target drug"), the characteristics of the hospital to be predicted (hereinafter referred to as the "target hospital"), and the period to be predicted (hereinafter referred to as the "target period"), and inputs this information into the central server 8. Note that information related to the type of target drug is referred to as target drug type information D9-1, information related to the characteristics of the target hospital is referred to as target hospital characteristic information D9-2, and information related to the period to be predicted is referred to as target period information D9-3.

The input method is to have information entered at a specific terminal be sent to the central server 8 via the communication network N, and then have the central server 8 receive it via the communication unit 83.

In this case, the target hospital characteristic information D9-2 uses information corresponding to the hospital characteristic information D8-3 included in the training data D8 used when generating the prediction model 823 in step S5-1. For example, if the hospital characteristic information D8-3 includes information related to the size of the hospital and the types of illnesses it treats, the target hospital characteristic information D9-2 should also include information related to the size of the target hospital and the types of illnesses it treats.

The target period related to the target period information D9-3 is set to the interval at which pharmaceutical inventory is replenished at the target hospital's pharmaceutical shop S (for example, one month if replenishment occurs monthly).

The target drugs, target hospitals, and target period may be set by the target hospitals themselves, and the data entered using the inventory management device 1 may be sent to the central server 8 via the wholesaler server 7.

When the above information is input, the control unit 81 of the central server 8 inputs input data D9, which includes at least target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3, into the prediction model 823.

Furthermore, if the information linked to the drug usage quantity information D8-1 in the training data D8 used to generate the prediction model 823 in step S5-1 includes information other than drug type information D8-2 and hospital characteristic information D8-3, it is preferable that the input data D9 input to the prediction model 823 in this step also include information corresponding to this.

For example, if the teacher data D8 includes seasonal information D8-4, it is preferable that the input data D9 also includes target period seasonal information D9-4, which is information regarding the season to which the target period belongs, in addition to target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3.
As the target period seasonal information D9-4, information regarding which period the target period belongs to (if it spans two periods, the longer period should be used) can be used so as to match the seasonal information D8-4 contained in the teacher data D8, for example, by defining March to May as spring, June to August as summer, September to November as autumn, and December to February as winter.

Furthermore, for example, if the teacher data D8 includes patient characteristic information D8-5, it is preferable that the input data D9 also includes target hospital patient characteristic information D9-5, which is information relating to the characteristics of patients actually hospitalized at the target hospital, in addition to target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3.
As the target hospital patient characteristic information D9-5, for example, information regarding the number of patients hospitalized at the target hospital for each symptom category (for example, dividing into injuries and illnesses, and then further dividing into specific body parts to define categories, and the number of patients in each category) can be used so as to match the patient characteristic information D8-5 included in the training data D8.

Furthermore, for example, if the teacher data D8 includes hospital location information D8-6, it is preferable that the input data D9 also includes target hospital location information D9-6, which is information related to the location of the target hospital, in addition to target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3.
The target hospital location information D9-6 can be information regarding which area the target hospital is located in, for example, within an area divided by prefecture, city, region, etc., so that it matches the hospital location information D8-6 included in the teacher data D8.

In the central server 8, the control unit 81 inputs this input data D9 into the prediction model 823 and outputs output data D10, which is information related to the predicted number of target drugs to be used in the target hospital during the target period, corresponding to the input data D9. This makes it possible to predict the number of target drugs to be used in the target hospital during the target period, as set by the administrator of the system, as described above.

[(3) Step S5-3: Prediction of Optimal Inventory Quantity]
Next, the flow of the process performed by the central server 8 when predicting the stock quantity of the target drug to be secured at the target hospital will be described.

The number of pharmaceuticals in stock that should be kept at the hospital (here, the number of each type of pharmaceutical M in stock that should be kept at the pharmaceutical shop S located within the hospital) must not be less than the number of each pharmaceutical M used during the period between replenishment of pharmaceutical M stock at the pharmaceutical shop S (for example, one month if stock is replenished monthly), and from a cost perspective, it is preferable not to hold excess stock.

The control unit 81 therefore predicts the appropriate inventory quantity of the target drug to be secured at the target hospital by adding a preset value as a margin to the usage quantity of the drug related to the output data D10 predicted in step S5-2 (the usage quantity of the target drug at the target hospital during the target period) to ensure that the usage quantity of the drug does not exceed the inventory quantity.

In this case, it is preferable that the value to be added to the number of uses of the target drug be a value determined individually for each type (name) of drug M.

That is, for example, for a drug M that is used frequently (used in large quantities), there is generally a great need to have a surplus in stock, so a large value is set to be added to the number of uses of the target drug when predicting the appropriate inventory quantity. In contrast, for a drug M that is used infrequently (used in small quantities), there is generally little need to have a surplus in stock, so a small value is set to be added to the number of uses of the target drug when predicting the appropriate inventory quantity.

Furthermore, for example, for pharmaceuticals M with a long expiration date, even if a large amount of stock is secured, there is a small risk that the expiration date will be exceeded and the product will be discarded, so a large value is set to be added to the number of pharmaceuticals to be used when predicting the appropriate inventory quantity. In contrast, for pharmaceuticals M with a short expiration date, there is a large risk that the expiration date will be exceeded and the product will be discarded when a large amount of stock is secured, so a small value is set to be added to the number of pharmaceuticals to be used when predicting the appropriate inventory quantity.

Furthermore, for example, for pharmaceuticals M that may pose a risk to a patient's life if they run out, such as drugs that are essential for the survival of a particular patient or surgical instruments that are essential for a particular surgery, the risk to a patient's life if they run out is high, so a large value is set to be added to the number of pharmaceuticals to be used when predicting the appropriate inventory level. In contrast, for pharmaceuticals M that are unlikely to pose a risk to a patient's life if they run out, a small value is set to be added to the number of pharmaceuticals to be used when predicting the appropriate inventory level, so the risk to a patient's life if they run out is low.

In this case, the target hospital may be allowed to set the value to be added to the number of uses of the target drug.
For example, at each hospital, a designated person in charge inputs the desired values for each type of medicine M in advance using the item management device 1, and the input values are acquired by the central server 8 and stored in the memory unit 82 for each hospital.In step S5-3, the values related to the hospital and medicine that match the target hospital and target medicine are read from the memory unit 82, and the read values are added to the number of uses related to the output data D10 output in step S5-2.

The control unit 81 stores optimum inventory quantity information D11, which is information related to the inventory quantity of the target drug that should be secured at the target hospital as predicted above, in the optimum inventory quantity information storage unit 824 of the storage unit 82, and then transmits the information to, for example, the product management device 1 used by the target hospital. This allows the hospital to order the target drug from the wholesaler in accordance with the optimum inventory quantity information D11.
In addition, for example, by sending the information to a wholesaler server 7 used by a wholesaler that handles the target pharmaceuticals, the wholesaler may be able to stock the pharmaceuticals in accordance with the appropriate inventory quantity information D11 at a pharmaceutical shop S located at the target hospital.

[Third explanation of effect]
Next, the effects of the item management system 100 according to this embodiment will be described.

According to the inventory management system 100 of this embodiment, a prediction model 823 that predicts the number of pharmaceuticals used is generated in the central server 8, and the generated prediction model 823 is used to predict the number of pharmaceuticals used, which is the target pharmaceutical, at the target hospital, which is the hospital that is the subject of the prediction, during the target period, which is the period that is the subject of the prediction.A preset value is then added to the predicted number of pharmaceuticals used, thereby predicting the inventory number of the target pharmaceutical that should be secured at the target hospital.
This makes it possible to predict the inventory quantity of target drugs that should be secured at target hospitals, with a certain margin over the number of target drugs expected to be used at the target hospitals during the target period. This reduces the risk that the inventory quantity of drugs will fall below the number of drugs to be used, while predicting an inventory quantity that is unlikely to be excessive, making it possible to predict the appropriate inventory quantity of drugs that should be secured at hospitals.

Furthermore, the training data D8 used when generating the prediction model 823 on the central server 8 in step S5-1 is data linking drug type information D8-2, hospital characteristic information D8-3, and drug usage quantity information D8-1, and the input data D9 input to the prediction model 823 when predicting usage quantities in step S5-2 includes target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3, making it possible to predict appropriate inventory quantities depending on the type of drug and the characteristics of the hospital.

Furthermore, if the training data D8 used when generating the prediction model 823 in the central server 8 in step S5-1 is data linking seasonal information D8-4, in addition to drug type information D8-2 and hospital characteristics information D8-3, with drug usage quantity information D8-1, and the input data D9 input to the prediction model 823 when predicting usage quantities in step S5-2 includes target drug type information D9-1, target hospital characteristics information D9-2, and target period information D9-3, as well as target period seasonal information D9-4, it becomes possible to predict appropriate inventory quantities according to the season of the target period. In other words, it is expected that the number of pharmaceuticals used will fluctuate significantly from season to season, for example, for drugs for treating diseases whose prevalence varies seasonally (e.g., influenza, which is prevalent in winter). In this case, it becomes possible to predict more appropriate inventory quantities for pharmaceuticals with such large seasonal usage fluctuations.

Furthermore, if the training data D8 used when generating the prediction model 823 on the central server 8 in step S5-1 is data in which patient characteristic information D8-5, in addition to drug type information D8-2 and hospital characteristic information D8-3, is linked to drug usage quantity information D8-1, and if the input data D9 input to the prediction model 823 when predicting usage quantities in step S5-2 includes target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3, as well as target hospital patient characteristic information D9-5, it becomes possible to predict appropriate inventory quantities based on the characteristics of patients actually hospitalized at the target hospital.

Furthermore, if the training data D8 used when generating the prediction model 823 on the central server 8 in step S5-1 is data in which, in addition to drug type information D8-2 and hospital characteristic information D8-3, hospital location information D8-6 is linked to drug usage quantity information D8-1, and if the input data D9 input to the prediction model 823 when predicting usage quantities in step S5-2 includes target drug type information D9-1, target hospital characteristic information D9-2, and target period information D9-3, as well as target hospital location information D9-6, it becomes possible to predict appropriate inventory quantities depending on the location of the target hospital.

Furthermore, when predicting the optimum inventory quantity in step S5-3, if the value to be added to the number of pharmaceuticals used related to the output data D10 output from the prediction model 823 is a value set for each type of pharmaceutical, it is possible to predict an inventory quantity for each type of pharmaceutical with an appropriate margin relative to the number of pharmaceuticals used.

Furthermore, in predicting the optimum inventory quantity in step S5-3, if the value to be added to the usage quantity of the drug related to the output data D10 output from the prediction model 823 is a value set by the target hospital, it is possible to reflect the hospital's requests and predict an inventory quantity with an appropriate margin relative to the usage quantity in accordance with the hospital's requests.

Furthermore, the processing from steps S1 to S3 can achieve the following effects:

In other words, the item management device 1 acquires the drug identification information D2 stored in the drug RFID tag 6 attached to the drug M located in the drug shop S and the staff identification information D1 stored in the staff RFID tag 5 carried by the staff member P who takes the drug M out of the drug shop S, links these together, and stores them in the entering staff member/taken out drug information storage unit 123 of the storage unit 12. This enables the item management device 1 to not only record the drug M taken out of the drug shop S and manage the inventory quantity, but also record information regarding which staff member P took out the drug M.

In addition, the item management device 1 acquires the staff identification information D1 read from the staff RFID tag 5 by the entry processing RFID reader 2 installed at the entrance of the pharmaceutical shop S, making it possible to acquire the staff identification information D1 of the staff P entering the pharmaceutical shop S at the time the staff P enters.

In addition, the item management device 1 acquires the drug identification information D2 read from the drug RFID tag 6 by the takeout processing RFID reader 3 installed in the drug shop S, and the staff identification information D1 read from the staff RFID tag 5 by the takeout processing RFID reader 3 installed in the drug shop S. This makes it possible to acquire the drug identification information D2 for the drug M being taken out and the staff identification information D1 for the staff P taking out the drug M at the time when the staff member P performs the takeout processing in step S2, that is, at the time when the staff member RFID tag 5 carried by the staff member P and the drug RFID tag 6 attached to the drug M being taken out are read by the takeout processing RFID reader 3.

In addition, the item management device 1 acquires drug identification information D2 read from the drug RFID tag 6 by the exit processing RFID reader 4 installed at the exit of the drug shop S, and staff identification information D1 read from the staff RFID tag 5 by the exit processing RFID reader 4 installed at the exit of the drug shop S, so that at the time the staff member P leaves the drug shop S, it is possible to acquire drug identification information D2 relating to the drug M being taken out and staff identification information D1 relating to the staff member P who takes out the drug M.
This allows the item management device 1 to obtain pharmaceutical identification information D2 relating to pharmaceuticals M taken out of the pharmaceutical shop S without being held over the sheet antenna 32 of the RFID reader 3 for take-out processing. Therefore, in the case of an emergency, such as a sudden change in a patient's condition or during surgery, it is possible to prevent a record of the pharmaceuticals M being taken out by staff P without holding the pharmaceuticals over the sheet antenna 32 of the RFID reader 3 for take-out processing from being lost.

Furthermore, when the item management device 1 acquires the drug identification information D2 read from the drug RFID tag 6 by the exit processing RFID reader 4 installed at the exit of the drug shop S when the staff member P leaves the drug shop S in step S3, it determines whether matching drug identification information D2 has also been acquired from the take-out processing RFID reader 3 in step S2, thereby making it possible to identify the drug M taken out of the drug shop S by the staff member P without having to hold it over the sheet antenna 32 of the take-out processing RFID reader 3.

[Fourth Modification]
Next, a modified example of the item management system 100 according to this embodiment will be described.

[1. Variation 1: Generation of a prediction model for each wholesaler]
In the explanation of the above operation, in step S5-1, the teacher data D8 is created from the aggregation result information D7 acquired from all wholesaler servers 7, and a prediction model 823 is generated using data related to all hospitals that use this system. However, instead, the teacher data D8 may be created only from the aggregation result information D7 acquired from a specific wholesaler server 7, and machine learning may be performed to generate a prediction model 823 individually for each wholesaler using only data related to hospitals to which a specific wholesaler wholesales pharmaceutical M (hospitals to which a specific wholesaler stocks pharmaceutical M at pharmaceutical shops S).
In this case, when predicting the number of uses in step S5-2, the number of uses of the drug will be predicted using a prediction model 823 generated using only data on wholesalers that supply drug M to the target hospital.

In this case, the processing of step S5 may be performed not by the central server 8 but by the wholesaler server 7 provided for each wholesaler.

While this reduces the amount of information that can be used as training data D8 when generating each prediction model 823, it enables more accurate predictions of drug usage in cases where, for example, trends in drug usage at hospitals differ between wholesalers.

[2. Modification 2: Generation of Prediction Model for Each Hospital]
In the explanation of the above operation, in step S5-1, the teacher data D8 is created from the aggregation result information D7 obtained from all wholesaler servers 7, and a prediction model 823 is generated using data related to all hospitals that use this system. However, instead, the teacher data D8 may be created only from the aggregation result information D7 related to a specific hospital, and machine learning may be performed to generate a prediction model 823 individually for each hospital using only data related to the specific hospital.
In this case, the prediction of the number of uses of the medicine in step S5-2 will also be performed using the prediction model 823 generated using only data related to the target hospital.

In this case, only data related to a single hospital is used to generate the prediction model 823 in step S5-1, and the prediction of the number of pharmaceuticals used in step S5-2 is performed using the prediction model 823 generated only from data related to the target hospital. Therefore, it is confirmed that the information related to the hospital's characteristics matches for all data contained in the training data D8 used to generate the prediction model and the input data D9 used to predict the number of pharmaceuticals used.

In this case, therefore, the training data D8 used in step S5-1 only needs to include at least drug type information D8-2 as information linked to drug usage quantity information D8-1, and does not need to include hospital characteristic information D8-3. Similarly, the input data D9 used in step S5-2 only needs to include at least target drug type information D9-1 and target period information D9-3, and does not need to include target hospital characteristic information D9-2.

In this case, the processing of step S5 may be performed not by the central server 8 but by the wholesaler server 7 provided for each wholesaler, or by the item management device 1 provided for each hospital.

As a result, predictions of the number of pharmaceutical uses will be made using prediction model 823 generated using only data related to the target hospital. Therefore, while versatility will be reduced as it can only be used for hospitals that have individually generated prediction model 823 in advance, the accuracy of predictions of the number of pharmaceutical uses can be further improved for hospitals that have generated prediction model 823.

[3. Modification 3: Generation of Prediction Model on Separate Server]
In the above description of the operation, the central server 8 uses the prediction model 823 it generates itself in step S5-1 to predict the number of pharmaceuticals used in step S5-2. However, it is also possible to provide a separate server for generating the prediction model, and then generate the prediction model in step S5-1 on that server.
In this case, the server transmits the generated prediction model 823 to the central server 8 via the communication network N, and the central server 8 acquires the prediction model 823 by receiving it via the communication unit 83, stores it in the memory unit 82, and then uses it to predict the number of pharmaceuticals used in step S5-2.

[4. Variation 4: Alert when take-out processing is incomplete]
When staff member P leaves the pharmaceutical shop S (step S3), the item management device 1 receives the staff member identification information D1 and pharmaceutical identification information D2 read by the exit processing RFID reader 4. If matching information is not stored in the entering staff member/taken-out pharmaceutical information storage unit 123 of the storage unit 12, that is, if staff member P takes pharmaceutical M out of the pharmaceutical shop S in step S2 without holding the staff RFID tag 5 and pharmaceutical RFID tag 6 over the sheet antenna 32 of the take-out processing RFID reader 3, the item management device 1 simply records the information in steps S3-5 and does not issue any warning.

In this regard, if the pharmaceutical shop S is located within a hospital and it is unlikely that an outsider will break in, it is sufficient to simply keep a record as described above. However, if the possibility of an outside intrusion cannot be ruled out, and it is necessary to prevent the removal of pharmaceuticals M that have not been processed for removal, a warning may be issued by a specified means.
For example, in step S3-3, if the control unit 11 determines that information matching the pharmaceutical identification information D2 received in step S3-2 is not stored in the entering staff/taken-out pharmaceutical information memory unit 123 of the memory unit 12, it may emit a predetermined alarm sound from an audio generating means such as a buzzer connected to the inventory management device 1, or it may send a predetermined message from the inventory management device 1 to a terminal used by a predetermined security officer, etc., warning that pharmaceutical M has been taken out without undergoing take-out processing.

[5. Variation 5: Installation of gates at entrances and exits]
In the above description of the configuration and operation, we have explained the case where no gate is installed at the entrance/exit E of the pharmaceutical shop S, but it is also possible to install a gate connected to the item management device 1 via a communication cable C at the entrance/exit E of the pharmaceutical shop S to prevent entry or exit without following proper procedures.
In this case, for example, if someone attempts to enter through the entrance/exit E without holding the staff RFID tag 5 over the sheet antenna 22 of the entry processing RFID reader 2, such a gate can be closed to prevent people without the staff RFID tag 5 from entering the pharmaceutical shop S.
Furthermore, for example, if someone attempts to leave the store through the entrance/exit E with a pharmaceutical product M without holding the pharmaceutical product RFID tag 6 over the sheet antenna 32 of the RFID reader 3 for removal processing, closing such a gate can prevent the pharmaceutical product M from being removed from the pharmaceutical shop S before the removal processing described in step S2 is completed.

[6. Variation 6: Entry processing using AI camera]
In the explanation of the above operation, in the entry processing of step S1, the employee P holds up the employee RFID tag 5 that he or she carries to the entry processing RFID reader 2 provided at the entrance/exit E of the pharmaceutical shop S, which reads the employee identification information D1 and sends it to the goods management device 1, thereby recording information related to the employee P who has entered the pharmaceutical shop S in the goods management device 1. However, instead of the entry processing using such an RFID tag and RFID reader, an AI camera capable of identifying the face of the employee P may be provided at the entrance/exit E of the pharmaceutical shop S, and the employee P entering the shop may be identified by facial recognition, so that information related to the employee P who has entered the pharmaceutical shop S is recorded in the goods management device 1.
This allows staff member P to enter the medicine shop S without having to hold up the staff member RFID tag 5 over the entrance processing RFID reader 2.

In addition, by performing both this type of processing using an AI camera and the processing using the RFID tag and RFID reader described in step S1, it may be possible to more reliably obtain information about staff member P who enters pharmaceutical shop S.

In addition to facial authentication, staff member P entering the room may be identified using other biometric authentication methods such as fingerprints, voiceprints, and retina.

[7 Variation 7: Assistance in item search using AI speaker]
When there are many medicines M in the medicine shop S, it may be difficult for the staff P to find the desired medicine M. In preparation for such a case, an AI speaker may be used to assist the staff P in searching for the medicine M, in order to make it easier for the staff P to search for the medicine M.
Specifically, information regarding the location of medicines M in the medicine shop S is stored in the medicine database 122 provided in the memory unit 12 of the goods management device 1, and an AI speaker connected to the goods management device 1 is provided in the medicine shop S. Then, when the staff member P states the name of the desired medicine M into the microphone provided in the AI speaker, data related to the matching medicine M is searched for in the medicine database 122, and information regarding the location of the medicine is provided by voice.

[8. Modification 8: Omission of Take-Out Processing]
In the above description of the operation, we have explained the case where staff identification information D1 related to staff P who takes out the medicine M and medicine identification information D2 related to the medicine M are obtained in both the take-out process in step S2 and the exit process in step S3.
In this regard, if the accuracy of reading the staff identification information D1 and drug identification information D2 by the RFID reader 4 for exit processing during exit processing is sufficient, it is possible to configure the system so that the staff identification information D1 related to the staff P who takes out the drug M and the drug identification information D2 related to the drug M are obtained simply by obtaining information when the staff P leaves the room using the RFID reader 4 for exit processing, without having to provide the RFID reader 3 for takeout processing.
However, in order to reduce the possibility of not being able to acquire this information, it is preferable to acquire the information twice, in both the take-out process in step S2 and the exit process in step S3, as described above.

[9 Variation 9: Execution of warehousing work]
When a pharmaceutical wholesaler delivers pharmaceuticals M to a pharmaceutical shop S, the pharmaceutical RFID tag 6 attached to the pharmaceutical M may be held over the sheet antenna 32 of the RFID reader 3 for removal processing, and only the information relating to the item from which pharmaceutical identification information D2 has been obtained may be stored in the pharmaceutical database 122 of the memory unit 12 of the item management device 1.
This reduces the risk that information relating to medicines that have not actually been delivered will be stored in the medicine database 122 due to input errors or the like.
As described above, if the ownership of the medicines M in the medicine shop S is not transferred to the hospital where the medicine shop S is located but remains with the wholesaler, then such a stocking operation is not necessary, since there is usually no problem even if information about medicines that have not actually been delivered is stored in the medicine database 122. However, for items such as medicines and surgical instruments, the shortage of which could be expected to pose a risk to human life, it is desirable to carry out such processing in order to prevent inventory shortages.

[10 Variation 10: Analysis of Drug Usage Trends]
In the item management device 1, the control unit 11 may classify (analyze trends) information related to the taking out of pharmaceuticals M based on information related to the taken out pharmaceuticals M stored in the entering staff/taken out pharmaceuticals information memory unit 123 of the memory unit 12 and information related to the staff member P who took it out.
For example, by classifying information by employee P's affiliation (e.g., by hospital department), it is possible to obtain information about the drug usage trends (which drugs are used most frequently) for each employee P's affiliation.

[11 Variation 11: Application to places other than pharmaceutical shops]
In step S5 of the above operation description, we have explained the case of predicting the appropriate inventory quantity for pharmaceuticals M to be placed in the pharmaceutical shop S, but the application of the appropriate inventory quantity prediction process in step S5 is not limited to the case of managing the inventory quantity of pharmaceuticals that are secured at hospitals by being placed in the pharmaceutical shop S.
For example, when a wholesaler sells and delivers pharmaceuticals directly to a hospital without going through Pharmaceutical Shop S, the hospital may go through a process similar to that described in step S5 when predicting the appropriate inventory of pharmaceuticals to be kept in a warehouse or the like within the hospital.

100 Item management system 8 General server (medicine inventory management device)
81 Control unit (prediction model generation/acquisition means, usage quantity prediction means, optimum inventory quantity prediction means)
82 Storage unit 823 Prediction model 83 Communication unit (prediction model generation/acquisition means)
D8 Training data D8-1 Drug usage quantity information D8-2 Drug type information D8-3 Hospital characteristic information D8-4 Season information D8-5 Patient characteristic information D8-6 Hospital location information D9 Input data D9-1 Target drug type information D9-2 Target hospital characteristic information D9-3 Target period information D9-4 Target period seasonal information D9-5 Target hospital patient characteristic information D9-6 Target hospital location information D10 Output data D11 Appropriate inventory quantity information

Claims (7)

A prediction model generation/acquisition means for generating or acquiring a prediction model that predicts the number of pharmaceuticals used in a hospital;
a usage quantity prediction means for predicting the usage quantity of a target drug, which is a drug that is the subject of prediction, at a target hospital, which is a hospital that is the subject of prediction, during a target period, which is a period that is the subject of prediction, using the prediction model;
an appropriate inventory quantity prediction means for predicting the inventory quantity of the target drug that should be secured in the target hospital by adding a preset value to the usage quantity of the target drug predicted by the usage quantity prediction means;
Equipped with
The prediction model generation/acquisition means generates the prediction model using information linking at least information related to the type of drug and information related to the characteristics of the hospital with information related to the number of drugs used in a predetermined period, or acquires the prediction model generated using information linking at least information related to the type of drug and information related to the characteristics of the hospital with information related to the number of drugs used in a predetermined period,
The pharmaceutical inventory management device is characterized in that the usage quantity prediction means inputs at least information related to the type of target pharmaceutical, information related to the characteristics of the target hospital, information related to the target period, and information related to the characteristics of patients hospitalized at the target hospital into the prediction model, and predicts the usage quantity of the target pharmaceutical at the target hospital during the target period . A prediction model generation/acquisition means for generating or acquiring a prediction model that predicts the number of pharmaceuticals used in a hospital;
a usage quantity prediction means for predicting the usage quantity of a target drug, which is a drug that is the subject of prediction, at a target hospital, which is a hospital that is the subject of prediction, during a target period, which is a period that is the subject of prediction, using the prediction model;
an appropriate inventory quantity prediction means for predicting the inventory quantity of the target drug that should be secured in the target hospital by adding a preset value to the usage quantity of the target drug predicted by the usage quantity prediction means;
Equipped with
The prediction model generation/acquisition means generates the prediction model using information linking at least information related to the type of drug and information related to the number of drugs used in the target hospital for a predetermined period of time, or acquires the prediction model generated using information linking at least information related to the type of drug and information related to the number of drugs used in the target hospital for a predetermined period of time,
The pharmaceutical inventory management device is characterized in that the usage quantity prediction means inputs at least information regarding the type of target pharmaceutical, information regarding the target period, and information regarding the characteristics of patients hospitalized at the target hospital into the prediction model, and predicts the usage quantity of the target pharmaceutical at the target hospital during the target period . The pharmaceutical inventory management device described in claim 1 or 2, characterized in that the usage quantity prediction means further inputs information related to the season to which the target period belongs into the prediction model and predicts the usage quantity of the target pharmaceutical at the target hospital during the target period . The pharmaceutical inventory management device described in any one of claims 1 to 3 , characterized in that the usage quantity prediction means further inputs information related to the location of the target hospital into the prediction model and predicts the usage quantity of the target pharmaceutical at the target hospital during the target period. The pharmaceutical inventory management device of any one of claims 1 to 4, characterized in that the predictive model generation/acquisition means generates the predictive model by machine learning using data linked at least to information related to the type of pharmaceutical and information related to the number of pharmaceuticals used as training data, or acquires the predictive model generated by machine learning using data linked at least to information related to the type of pharmaceutical and information related to the number of pharmaceuticals used as training data. A pharmaceutical inventory management device as described in any one of claims 1 to 5, characterized in that the value added by the optimum inventory quantity prediction means is a value set for each type of pharmaceutical, and the value added is set to be larger for pharmaceuticals that are used frequently, pharmaceuticals with long expiration dates, and pharmaceuticals whose shortage may pose a risk to the patient's life than for other pharmaceuticals . 7. The pharmaceutical inventory management device according to claim 1 , wherein the value added by the optimum inventory quantity prediction means is a value set by the target hospital.