Medicine management device, medicine management-purpose program, and learning model preparation device

Abstract

To provide an art that enables placing an order for required medicine even if a schedule of medically relevant activities such as an inspection and the like in a medical institution is not determined.SOLUTION: A medicine management device 11 comprises: a discriminator 13 that is constructed by information on a schedule of past medically relevant activities performed in medical institutions, and a learning model prepared by machine learning using teacher data including information on a kind and amount of medicine used in the medically relevant activity; a schedule input reception unit 142 that receives an input of information on a future schedule; and an order-placement information preparation unit 143 that inputs the schedule on the future schedule, and prepares order-placement information on medicine on the basis of the information on the kind and amount of the medicine to be output from the discriminator.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to techniques for managing pharmaceuticals, and in particular, to techniques for managing radiopharmaceuticals. [Background technology]

[0002] Nuclear medicine tests (also called radioisotope (RI) tests) are one of the tests performed at medical institutions. Nuclear medicine tests are used to examine the pathology of diseases in various parts of the body, such as the heart, brain, respiratory system, renal and urinary system, and bones and joints. In nuclear medicine tests, a drug (radiopharmaceutical) containing a small amount of radioactive isotope is administered to the area to be examined, and the radiation emitted from it is measured.

[0003] Because the radiation emitted from the radioisotopes contained in radiopharmaceuticals decreases over time, they cannot be stored for long periods at medical institutions. For this reason, traditionally, at medical institutions, technicians or other personnel would check reservations for nuclear medicine examinations accepted by each department just before the examination date (for example, the day before), specify the type and amount of pharmaceuticals to be used for each examination, and place orders with the manufacturer.

[0004] As mentioned above, the target areas of nuclear medicine examinations vary, and the type of radiopharmaceutical used varies depending on the examination area. In addition, the dosage varies depending on the age and physique of the patient. For this reason, it is time-consuming and laborious for the person in charge to check the details of the nuclear medicine examination scheduled at the medical institution and decide the type and amount of radiopharmaceutical to order.

[0005] Patent Document 1 describes an ordering support device for radiopharmaceuticals. When a reservation for a nuclear medicine examination to be performed at a medical institution is confirmed, this ordering support device obtains reservation information including the date, time, and details. Then, by searching a product information table prepared in advance, the device identifies the type and amount of radiopharmaceutical corresponding to the obtained examination details, creates ordering information, and transmits it to the manufacturer. By using this ordering support device, the person in charge of ordering does not need to specify the type and amount of radiopharmaceutical, and radiopharmaceuticals can be ordered easily and without error. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2018-28517 A Summary of the Invention [Problem to be solved by the invention]

[0007] In the ordering support device described in Patent Document 1, the type and amount of radiopharmaceuticals to be ordered are determined based on reservation information for a nuclear medicine examination, and therefore the type and amount of radiopharmaceuticals to be ordered cannot be determined until the reservation for the nuclear medicine examination is confirmed. Since it takes a certain amount of time to manufacture and deliver radiopharmaceuticals, the order for the pharmaceuticals to be used in the nuclear medicine examination must be completed well before the nuclear medicine examination is performed. Therefore, even if there is a vacant slot for the nuclear medicine examination at the predetermined ordering time, reservations cannot be accepted after the ordering time, and there are cases where the slot for the nuclear medicine examination cannot be used effectively.

[0008] The problem to be solved by the present invention is to provide a technique that enables a necessary medicine to be ordered even if the schedule of medical procedures such as tests at a medical institution has not been confirmed. [Means for solving the problem]

[0009] The medicine management device according to the present invention, which solves the above problems, A classifier constructed by a learning model created by machine learning using training data including information on the schedule of past medical procedures performed at a medical institution and information on the types and amounts of medicines used in the medical procedures; a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on information on the type and amount of the drug output from the classifier; The present invention is characterized by comprising:

[0010] In the drug management device according to the present invention, a classifier is constructed using a learning model created by machine learning using data including information on the schedule of medical procedures previously performed at a medical institution and information on the type and amount of medicine used in the medical procedures, i.e., the actual use of medicines at the medical institution as training data. More specifically, this machine learning inputs the schedule of medical procedures and outputs the type and amount of medicine. The medical procedures referred to here may include medical procedures such as tests using medicines, as well as, for example, maintenance of medical equipment.

[0011] The work schedules of doctors and technicians (collectively referred to as "doctors, etc.") at medical institutions are often patterned on a regular cycle (for example, on a weekly basis). Furthermore, the content of medical procedures and the attributes of patients are often typical for each doctor, etc. Furthermore, the cycle for maintenance of medical equipment is often also fixed (for example, every six months or once a year). Therefore, there is a correlation between the schedule of past medical procedures performed at a medical institution and the type and amount of medicine used in the medical procedures. In the medicine management device according to the present invention, a learning model is created by machine learning such correlations, and a classifier is constructed.

[0012] In the drug management device according to the present invention, first, a user inputs a future schedule to the schedule input reception unit. The future schedule here refers to a schedule for ordering drugs that are expected to be used in a medical-related procedure. The order information creation unit inputs the input schedule to the above-mentioned identifier, and creates drug order information based on the information on the type and amount of the drug output from the identifier. In this way, the drug management device according to the present invention creates drug order information without using reservation information for the medical-related procedure, so that necessary drugs can be ordered even if the schedule for the medical-related procedure is not yet confirmed. In addition, even for examination slots for which reservations have not been accepted at this point, order information including the type and amount of drugs required for the expected medical-related procedure is created based on the past performance of the medical-related procedure, so that the slots for nuclear medicine examinations can be effectively used.

[0013] In the medicine management device according to the present invention, The information on the schedule of the past medical care procedure includes information on the day of the week, The schedule input receiving unit receives input of schedule information including information on days of the week. It is preferable.

[0014] In medical institutions, the work schedules of doctors and other medical personnel are often patterned on a weekly basis. Therefore, in many cases, the types and amounts of medicines used in medical procedures also have a pattern that is synchronized with this. Therefore, by including information on the days of the week in the schedule information of past medical procedures used as training data, machine learning that includes learning of this pattern can be performed. In addition, by inputting schedule information including the days of the week into the schedule input reception unit, order information can be created with higher accuracy.

[0015] The medicine management device according to the present invention further comprises: A display unit; a display control unit that causes the order information to be displayed on the display unit; an order information edit receiving unit that receives edits of information on the type and amount of medicine included in the order information; It is preferable to have

[0016] In the pharmaceutical management device of the above embodiment, when a reservation for an irregular medical procedure is accepted or when irregular inspections of medical equipment are performed, the user can appropriately change the type and amount of pharmaceuticals required for those medical procedures.

[0017] A medicine management program according to another aspect of the present invention includes: A computer in which a classifier is stored, the classifier being constructed by a learning model created by machine learning using training data including information on the schedule of past medical-related procedures performed at the medical institution and information on the type and amount of medicine used in the medical-related procedures, inputting information on the schedule of the medical-related procedures and outputting information on the type and amount of the medicine, a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on the information on the type and amount of the drug output from the classifier; The present invention is characterized in that it operates as

[0018] According to yet another aspect of the present invention, there is provided a learning model creation device, A storage unit that stores information on the schedule of past medical procedures performed at a medical institution and information on the types and amounts of medicines used in the medical procedures; a learning model creation unit that creates a learning model by machine learning using information on the schedule of the medical care-related procedure as an input and information on the type and amount of the medicine as an output; The present invention is characterized by comprising: Effect of the Invention

[0019] By using the medicine management device according to the present invention, it is possible to order necessary medicines even if the schedule for medical procedures such as examinations at medical institutions has not been confirmed. [Brief description of the drawings]

[0020] [Figure 1]1 is a diagram showing the overall configuration of a medicine management system including a medical institution server which is an embodiment of a medicine management device according to the present invention. [Diagram 2] FIG. 2 is a diagram showing the main configuration of a medical institution server according to the embodiment. [Diagram 3] FIG. 2 is a diagram showing the main configuration of the department terminal according to the embodiment. [Figure 4] 13 shows an example of a screen display (trend) in this embodiment. [Diagram 5] 4 shows an example of a screen display (order) in this embodiment. [Figure 6] 13 shows an example of a screen display (order completed) in the present embodiment. [Figure 7] 4 shows an example of a screen display (history) in the present embodiment. [Figure 8] 13 shows an example of a screen display (analysis) in this embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] A medical institution server, which is an embodiment of a drug management device and a learning model creation device according to the present invention, and an embodiment of a drug management program according to the present invention will be described below with reference to the drawings. The drug management device of this embodiment is used to manage radiopharmaceuticals used in nuclear medicine examinations, etc.

[0022] 1 shows the overall configuration of a medicine management system 1 according to this embodiment. The medicine management system 1 according to this embodiment is roughly composed of a medical institution system 10 and a contractor device 31.

[0023] The medical institution system 10 includes a medical institution server 11 and a medical department terminal 21 owned by each medical department. The medical institution server 11 and the medical department terminal 21 are connected to each other via a network within the medical institution. The medical institution server 11 and the contractor device 31 are connected to each other via the Internet. In FIG. 1, four terminals (A medical department terminal 21a, B medical department terminal 21b, C medical department terminal 21c, and D medical department terminal 21d) are shown as the medical department terminals 21, but the number of medical department terminals 21 can be changed arbitrarily. The medical institution server 11 of this embodiment has a function as a medicine management device in the present invention and a function as a learning model creation device.

[0024] 2 shows the main configuration of the medical institution server 11. The medical institution server 11 has a storage unit 12 and a classifier 13. The storage unit 12 is provided with a patient information storage unit 121, a medical worker information storage unit 122, a performance information storage unit 123, a reservation information storage unit 124, and a medical-related procedure database (DB) 125.

[0025] The patient information storage unit 121 stores information about patients who have previously received medical treatment such as examinations and tests at the medical institution. The patient information may include identification information (e.g., a patient ID) for identifying each patient, as well as the patient's name, address, telephone number, and other information.

[0026] The medical staff information storage unit 122 stores information on medical staff (doctors, engineers, nurses, etc.) working at the medical institution. The medical staff information may include identification information (e.g., staff ID) for identifying each medical staff, login password, attributes (doctor, engineer, nurse, etc.), name, address, telephone number, department of work, work schedule, etc. In addition, the identification information is associated with various usage authority information including the permission or disapproval of pharmaceutical ordering operations, which will be described later. Note that this information may also include information on those who have worked at the medical institution in the past, information on those who are scheduled to work at the medical institution in the future, and information on medical staff (doctors, engineers, nurses, etc.) who belong to institutions other than the medical institution in question (e.g., another medical institution or university).

[0027] The performance information storage unit 123 stores performance information of medical-related procedures performed in the past in each medical department of the medical institution. This information includes information such as the date, day of the week, and time when the medical-related procedure was performed, the medical department in which the medical-related procedure was performed, identification information of the medical worker in charge, the type of medical-related procedure, the equipment used, the type and amount of medicine used in the medical-related procedure, and identification information of the patient who was the subject of the medical-related procedure. Note that the medical-related procedure referred to here includes not only the medical procedure itself such as an examination, but also maintenance such as inspection of medical equipment.

[0028] The reservation information storage unit 124 stores reservation information for medical-related procedures accepted by each department of the medical institution. This information includes the date and time of the reservation, the department that accepted the reservation, identification information of the medical staff in charge of the medical-related procedure, the type of medical-related procedure, the equipment used, and identification information of the patient for the medical-related procedure. Here, too, reservations for medical-related procedures include reservations for the medical procedures themselves, such as examinations, as well as schedules for maintenance, such as inspections of medical equipment.

[0029] The medical-related procedure database 125 is a database that associates the contents of various medical-related procedures, such as tests performed in each department of the medical institution and maintenance of medical equipment, with information on medicines required for each medical-related procedure. This medical-related procedure database 125 not only stores information that associates the type of medical-related procedure with information on medicines, but also stores information on the type and amount of medicine used for the same test, for example, if the type and amount of medicine used differ depending on the sex or weight of the patient (information that associates the combination of the type of medical-related procedure and the attributes of the patient with the type and amount of medicine).

[0030] The classifier 13 is created by the learning model creation unit 141 described below, and is configured to output information on the type and amount of medicine corresponding to a future schedule when the schedule is input.

[0031] The medical institution server 11 also includes, as functional blocks, a learning model creation unit 141, a schedule input reception unit 142, an order information creation unit 143, a display control unit 144, an order information edit reception unit 145, an order execution unit 146, an order completion data acquisition unit 147, a history data acquisition unit 148, and an analysis processing unit 149. A general server computer or personal computer can be used as the medical institution server 11, and these functional blocks are realized by executing a medicine management program pre-installed in the computer with a processor. An input unit 18 including a mouse and a keyboard, and a display unit 19 consisting of a liquid crystal display or the like are connected to the medical institution server 11.

[0032] FIG. 3 shows the main components of the department terminal 21. The department terminal 21 includes a storage unit 22. The storage unit 22 includes a patient information storage unit 221 and a medical worker information storage unit 222. The patient information storage unit 221 stores information on patients in the department. The medical worker information storage unit 222 stores information on medical workers in charge of the department. The contents of the patient information and medical worker information stored therein are the same as those stored in the patient information storage unit 121 and the medical worker information storage unit 122. When a new patient is accepted in the department, the patient information is appropriately input and stored in the patient information storage unit 221. When a new medical worker is assigned to the department or when a medical worker resigns, the medical worker information is appropriately input or deleted, and the information in the medical worker information storage unit 222 is updated.

[0033] The department terminal 21 also includes, as functional blocks, a reservation information receiving unit 231 and an information transmitting unit 232. A general personal computer or tablet terminal can be used as the department terminal 21, and these functional blocks are realized by executing a program for the department terminal preinstalled in the computer on a processor. An input unit 28 including a mouse and keyboard, and a display unit 29 consisting of a liquid crystal display or the like are connected to the department terminal 21. When the department terminal 21 is a tablet terminal, a display unit 29 such as a liquid crystal display that also has the function of the input unit 28 is used.

[0034] The reservation information reception unit 231 receives a reservation for a medical-related procedure in the department. Specifically, the reservation information includes the date and time of the reservation, the department that received the reservation, the identification information of the medical worker in charge of the medical-related procedure, the type of medical-related procedure, the equipment used, and the identification information of the patient who is the target of the medical-related procedure. Here, the reservation for the medical-related procedure may include a reservation for the medical procedure itself such as an examination, as well as a schedule for maintenance such as inspection of the medical equipment. When receiving a reservation for a medical procedure such as an examination for a new patient, the input of information such as the name, address, and telephone number of the patient is also received. The information of the new patient received here is sequentially stored in the patient information storage unit 221. The reservation information reception unit 231 also receives an input of information on the results of the medical-related procedure received by the department (for example, whether or not the medical-related procedure was performed, information on the actual dosage of the medicine, etc.). The reservation information for the medical-related procedure received by the reservation information reception unit 231 and the information on the results are sequentially stored in the storage unit 22.

[0035] When information in the patient information storage unit 221 or the medical worker information storage unit 222 is updated, the information transmission unit 232 transmits the information together with information identifying the medical department to the medical institution server 11. When the medical institution server 11 receives the information, it sequentially reflects it in the patient information storage unit 121 and the medical worker information storage unit 122. Similarly, when a reservation for a medical-related procedure is accepted in the medical department, and when information on the results of the medical-related procedure is input, the information transmission unit 232 transmits the information to the medical institution server 11. The medical institution server 11 stores information on the accepted reservation in the reservation information storage unit 124, and stores information on the results of the medical-related procedure in the performance information storage unit 123.

[0036] The order-receiving device 31 is used by a party that receives an order for medicines from a medical institution and delivers them to the medical institution, such as a company that manufactures medicines (medicinal product manufacturing company) or a company that wholesales medicines manufactured by a pharmaceutical manufacturing company (medicinal product wholesale company). The order-receiving device 31 accepts order information for medicines manufactured by the company.

[0037] The flow of managing medicines in the medicine management system 1 of this embodiment will be described below.

[0038] In this embodiment, first, the classifier 13 used in the medical institution server 11 is created. When the user instructs the creation of the classifier 13, the learning model creation unit 141 reads out the performance information of the past medical-related procedures performed in each medical department of the medical institution from the performance information storage unit 123, and creates a learning model using the information as teacher data. As a specific method of supervised learning and an algorithm that is the basis of the learning model, various conventionally known methods can be used. In this embodiment, as the performance information, information on the date, day of the week, time period, medical department, medical worker in charge, type of medical-related procedure, age, sex, and weight of the patient, and type and amount of medicine used in the medical-related procedure are read out. Regarding the information on the medical worker, the identification information of the medical worker stored in the performance information storage unit 123 is used to read out necessary information (name and attributes of the medical worker (doctor, technician, etc.)) from the medical worker information storage unit 122, and regarding the information on the patient, the identification information of the patient stored in the performance information storage unit 123 is used to read out necessary information (age, sex, weight, etc.) from the patient information storage unit 121. Then, machine learning is performed, inputting schedule information and outputting information on the type and amount of medicine, to create a learning model.

[0039] In general, in medical institutions, the work schedules of medical practitioners such as doctors in each medical department are patterned. In particular, the working hours of medical practitioners such as doctors are often patterned on a weekly basis. In addition, the type of medical-related procedure performed by each doctor and the attributes of patients (age, sex, and weight) are often unique to each doctor. Furthermore, as stored in the medical-related procedure database 125, the type and amount of medicine used in the medical-related procedure are determined by the type of medical-related procedure and the attributes of the patient. In addition, maintenance of testing equipment that may be included in the medical-related procedure is usually performed at regular intervals (every six months or once a year). Therefore, there is a correlation between the schedule information included in the performance information of the medical-related procedure and the type and amount of medicine used in the medical-related procedure performed by each doctor. In this embodiment, a learning model is created by machine learning using the performance information of the medical-related procedure as training data. Then, using the created learning model, a classifier 13 is constructed that inputs schedule information and outputs information on the type and amount of medicine, and is installed in the medical institution server 11.

[0040] When a person who wishes to place an order for a medicine (an ordering person such as a technician) logs in using his / her own identification information and password on a specified authentication screen of the medical institution server 11 in which the identifier 13 is installed, the schedule input acceptance unit 142 checks the entered identification information and password against those stored in the medical worker information storage unit 122, and confirms whether the medical worker holding that identification information has been granted the authority to order medicines.

[0041] If it is confirmed that the identification information and the password match and that the user has the authority to order medicines, the schedule input accepting unit 142 displays a screen for ordering medicines on the display unit 19. An example of the screen displayed on the display unit 19 is shown in FIG.

[0042] 4, a menu and the current date are displayed at the top of the screen. The menu has five buttons: a trend button 411, an order button 412, an order button 413, a history button 414, and an analysis button 415. There is also a field 42 for displaying the current date and the currently logged-in user.

[0043] When the user presses the trend button 411, the schedule input acceptance unit 142 displays a date input field on the screen and allows the user to input the date of use of the drug to be ordered. The date input field is composed of a date display field 431 and a calendar button 432. Pressing the calendar button 432 displays a calendar in a separate window, and when a date is specified on the calendar, information on the specified date and day of the week is displayed in the date display field 431.

[0044] When the user inputs the date of use of the medicine, a start button 44 is displayed next to the date input field. When the user presses the start button 44, the order information creation unit 143 inputs the information of the date of use of the medicine (year, month, day, and day of the week) input by the user to the classifier 13. The information of the predicted use of the medicine output from the classifier 13 (information including the examination name, medical department, doctor in charge, medicine name, time period of use, patient weight, and number of medicines ordered) is displayed in a table format on the screen by the display control unit 144 (Table 45 in FIG. 4). As described above, the classifier 13 has learned the patterns and trends of the contents of the medical-related procedures and the types and amounts of medicines based on the past records of medical-related procedures performed at the medical institution and the records of medicines used in the medical-related procedures. Therefore, this Table 45 reflects information such as the contents of the medical-related procedures at the medical institution and the types and amounts of medicines estimated based on the past records.

[0045] In this manner, in this embodiment, since pharmaceutical order information can be created based on the information output by the identifier 13, necessary pharmaceuticals can be ordered even if the schedule for the medical procedure has not been confirmed.

[0046] The display control unit 144 displays a table 45 based on the information output from the classifier 13, and also displays a comparison button 46 and a preparation complete button 47 on the screen. If the user presses the preparation complete button 47 at this point, the order information creation unit 143 determines the information output from the classifier 13 (table 45) as planned order information.

[0047] On the other hand, when the user presses the comparison button 46, the order information editing reception unit 145 reads out the information of the reservation already accepted at that time from the reservation information storage unit 124. Then, it compares it with the information output from the discriminator 13. The comparison result is displayed on the screen in a table format by the display control unit 144 (Table 48 in FIG. 4). At this time, the order information editing reception unit 145 extracts the contents of the medical-related procedures (information such as the type of medical-related procedure and the attributes of the patient) from the reservation information that has already been accepted, and identifies the type and amount of medicine required for each of those medical-related procedures by referring to the medical-related procedure database 125. Then, if the type of medicine output from the discriminator 13 does not include the type of medicine required for the medical-related procedure for which the reservation has already been accepted, or if the amount of medicine is insufficient, a comparison result is created in which the type and / or amount of medicine is automatically added to satisfy them. The items and amounts added during comparison are displayed in a form different from the other items and amounts (such as coloring or blinking; hatching in FIG. 4).

[0048] In this manner, in this embodiment, since the ordering information can be corrected by comparing the information output by the identifier 13 with the information of medical procedures for which reservations have already been accepted, even if a reservation for an examination or the like that differs from the past patterns of medical procedures, such as a sudden examination, is accepted, the ordering information can be appropriately changed accordingly. Also, even for examination slots for which reservations have not yet been accepted at this point, the type and amount of pharmaceuticals required for the expected medical procedure are ordered based on the past performance of medical procedures, so that the slots for nuclear medicine examinations can be used effectively.

[0049] After the above comparison, when the user presses the preparation complete button 47, the order information creation unit 143 determines the information after the above comparison (the information revised based on the content of the medical-related procedure for which a reservation has already been accepted) as the planned order information.

[0050] After the order schedule information is determined, the user presses the order button 412, and the screen shown in FIG. 5 is displayed. FIG. 5 shows the content (same as Table 48) that is displayed when the above-mentioned comparison is performed. An approval button 51 is displayed on this screen. When the user presses the approval button 51, the order execution unit 146 transmits the information on the type and amount of the medicine displayed on the screen to the supplier device 31. This completes the ordering process for the medicine. When the supplier terminal 31 accepts an order from the medical institution server 11, it generates an order number for each medicine and transmits it to the medical institution server 11. When the medical institution server 11 receives the order number, the order execution unit 146 associates the order number with each medicine in the information on the type and amount of the ordered medicine (order history data) and stores it in the performance information storage unit 123.

[0051] When the user presses the order button 413, the screen shown in FIG. 6 is displayed. When the user presses the calendar button 62 on this screen, specifies a date from the calendar displayed in a separate window, and presses the decision button 63, the ordered data acquisition unit 147 reads out from the performance information storage unit 123 the order history data of the medicines scheduled for use on that date that are assigned an order number (orders completed by the manufacturing company, etc.; ordered data). Next, the display control unit 144 displays the date and day of the week specified by the user in the schedule display field 61, and also displays the ordered data of the medicines read out by the ordered data acquisition unit 147 in a table format (table 64 in FIG. 6). The display of the order history includes information on the type and amount of the medicine, the orderer, the order number, the order date and time, and the delivery date and time. By looking at this screen, the user can check the order status and delivery status of the medicines scheduled for use on that date. In addition, if there are medicines that are not displayed on this screen even though the ordering process has already been completed, it is possible that there may have been a transmission error or an order processing error on the part of the manufacturer, etc.

[0052] When the user presses the history button 414, the screen shown in FIG. 7 is displayed. This screen is provided with a search range specification field consisting of a start date input field 71 and an end date input field 72 for the search range. When the user inputs a date in each of these fields and presses the search button 73, the history data acquisition unit 148 reads out the ordered data acquisition unit 147 for the drug for the date of the search range input by the user from the performance information storage unit 123. This data is displayed on the screen by the display control unit 144. This list is displayed as data in a table format (table 76 in FIG. 7). The user can rearrange the data by clicking on the item name. In addition, by specifying an item in the item specification field 74 or specifying specific content in the input field 75 and then pressing the search button 73 (for example, specifying the item of the attending physician and searching by specifying physician A), it is possible to extract only the data of the specified content. In addition, by pressing the download button 77, the data displayed on the screen can be downloaded as a PDF file.

[0053] When the user presses the analysis button 415, the screen shown in FIG. 8 is displayed. This screen is provided with a start date input field 81 and an end date input field 82 for the analysis range. When the user inputs the date in each of these fields and presses the start button 83, the analysis processing unit 149 reads out the performance data of the medical-related procedures in the target range date input by the user from the performance information storage unit 123 and performs a predetermined statistical processing. The display control unit 144 displays the results of the statistical processing on the screen as a graph. The contents of the statistical processing can be arbitrarily set as necessary for the medical institution. In FIG. 8, as an example, the number of medical-related procedures per medical worker 84, the amount of medicine used per medical worker 85, and the number of medical-related procedures per medical department 86 are each shown as a pie chart or a bar graph.

[0054] In the above embodiment, the learning model creation unit 141 may update the classifier 13 at regular intervals or at a timing specified by the user by performing further machine learning on the installed classifier 13. For example, when a medical professional working at a medical institution retires or is newly appointed, the pattern of medical-related actions may change from past performance. In this way, by updating the classifier 13 by machine learning based on the latest performance data, it becomes possible to estimate more accurate drug ordering information based on the changed pattern even if a change occurs in the medical professional working at the medical institution.

[0055] The above embodiment is merely an example and can be modified as appropriate in accordance with the spirit of the present invention.

[0056] In the above embodiment, the order information edit receiving unit 145 reads out information on accepted reservations from the reservation information storage unit 124, compares it with the information output from the identifier 13, and automatically corrects the type and amount of the missing medicine, but the user may correct it himself. In that case, the order information edit receiving unit 145 may be configured to display on a screen a list of reservation information read out from the reservation information storage unit 124 and allow the user to edit the order information output from the identifier 13.

[0057] In the above embodiment, the medical institution server 11 performs the ordering process for pharmaceuticals, but the pharmaceutical order receiving party, such as a manufacturing company, may also perform the pharmaceutical order receiving process. In that case, the pharmaceutical management program in the above embodiment may be installed in the order receiving party device 31, and a classifier may be created and stored based on performance information provided by the medical institution. Also, the reservation information provided by the medical institution may be temporarily stored in the storage unit, and the data output from the classifier may be collated with the reservation information to create order information including information on the type and amount of pharmaceuticals. Also, preferably, the created order information may be sent by e-mail or the like to a person in charge at the medical institution, and the order may be confirmed after approval is obtained. Alternatively, a cloud server in which the pharmaceutical management program in the above embodiment is installed may be provided, and a configuration may be adopted in which it can be accessed by both the medical institution and the order receiving party, such as a manufacturing company.

[0058] In the above embodiment, it is assumed that one medical institution is provided with one medical institution server 11, but a common medical institution server may be provided for multiple medical institutions. In that case, identification information for identifying a medical institution may be stored in association with patient information, medical staff information, reservation information, etc., and a different identifier may be registered for each medical institution.

[0059] In the above embodiment, it is assumed that radiopharmaceuticals are to be managed, and the order information is created by specifying only one planned date of use of the pharmaceutical, but when managing general pharmaceuticals that can be managed for a certain period of time without deterioration, multiple days may be selected at the same time to create the order information. In that case, it is sufficient to configure the system so that the start date and end date of the order target period are input separately.

[0060] [Aspects] It will be apparent to those skilled in the art that the above-described exemplary embodiments are illustrative of the following aspects.

[0061] (Section 1) A medicine management device according to one aspect of the present invention includes: A classifier constructed by a learning model created by machine learning using training data including information on the schedule of past medical procedures performed at a medical institution and information on the types and amounts of medicines used in the medical procedures; a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on information on the type and amount of the drug output from the classifier; The present invention is characterized by comprising:

[0062] (Section 6) A medicine management program according to another aspect of the present invention includes: A computer in which a classifier constructed by a learning model created by machine learning using training data including information on the schedule of past medical procedures performed at the medical institution and information on the type and amount of medicine used in the medical procedures is stored, a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on the information on the type and amount of the drug output from the classifier; The present invention is characterized in that it operates as

[0063] (Section 7) According to yet another aspect of the present invention, there is provided a learning model creation device, A storage unit that stores information on the schedule of past medical procedures performed at a medical institution and information on the types and amounts of medicines used in the medical procedures; a learning model creation unit that creates a learning model by machine learning using information on the schedule of the medical care-related procedure as an input and information on the type and amount of the medicine as an output; The present invention is characterized by comprising:

[0064] In the pharmaceutical management device according to paragraph 1, the pharmaceutical management program according to paragraph 6, and the learning model creation device according to paragraph 7, a classifier is constructed using a learning model created by machine learning using data including information on the schedule of medical procedures previously performed at a medical institution and information on the type and amount of medicine used in the medical procedures, that is, the actual use of medicines at the medical institution as training data. More specifically, this machine learning takes the schedule of medical procedures as input and the type and amount of medicine as output. The medical procedures referred to here may include medical procedures such as tests using medicines, as well as, for example, maintenance of medical equipment.

[0065] The work schedules of doctors and technicians (collectively referred to as "doctors, etc.") at medical institutions are often patterned on a regular cycle (for example, on a weekly basis). In addition, the content of medical procedures and the attributes of patients are often standardized for each doctor, etc. Furthermore, the cycle for performing maintenance, etc. on medical equipment is often also fixed (for example, every six months or once a year). Therefore, there is a correlation between the schedule of past medical procedures performed at a medical institution and the type and amount of medicines used in those medical procedures. In the medicine management device related to paragraph 1, a learning model is created by machine learning such correlations, and a classifier is constructed.

[0066] In the pharmaceutical management device according to paragraph 1 and the pharmaceutical management program according to paragraph 6, first, a user inputs a future schedule to the schedule input receiving unit. The future schedule here refers to a schedule for which pharmaceuticals expected to be used in a medical-related procedure are to be ordered. The order information creating unit inputs the input schedule to the above-mentioned classifier, and creates pharmaceutical order information based on information on the type and amount of pharmaceuticals output from the classifier. In this way, in the pharmaceutical management device according to paragraph 1 and the pharmaceutical management program according to paragraph 6, pharmaceutical order information is created without using reservation information for a medical-related procedure, so that necessary pharmaceuticals can be ordered even if the schedule for the medical-related procedure has not been confirmed.

[0067] (Section 2) The drug management device according to paragraph 2 is a drug management device according to paragraph 1, The information on the schedule of the past medical care procedure includes information on the day of the week, The schedule input receiving unit receives input of schedule information including information on days of the week.

[0068] In medical institutions, the work schedules of doctors and other medical personnel are often patterned on a weekly basis. Therefore, in many cases, the types and amounts of medicines used in medical procedures also have a pattern that is synchronized with this. In the medicine management device according to paragraph 2, by including information on the days of the week in the schedule information of past medical procedures used as training data, machine learning including learning of this pattern can be performed. In addition, by inputting schedule information including the days of the week into the schedule input receiving unit, it is possible to create order information with higher accuracy.

[0069] (Section 3) The drug management device according to paragraph 3 is the drug management device according to paragraph 1 or 2, further comprising: A display unit; a display control unit that causes the order information to be displayed on the display unit; an order information edit receiving unit that receives edits of information on the type and amount of medicine included in the order information; Equipped with.

[0070] In the case of the medicine management device under paragraph 3, when a reservation for an irregular medical procedure is accepted or when irregular inspection of medical equipment is performed, the user can appropriately change the type and quantity of medicine required for that medical procedure.

[0071] (Section 4) The drug management device according to paragraph 4 further comprises the drug management device according to paragraph 3, a medical care-related procedure database that stores information on the types and amounts of medicines required when performing the medical care-related procedure; a reservation information storage unit in which reservation information for medical procedures accepted at the medical institution is stored; Equipped with The order information editing reception unit identifies the type and quantity of medicines required for the reserved medical-related procedure stored in the reservation information storage unit based on the medical-related procedure database, and if the type and quantity of medicines contained in the order information are insufficient, the order information is modified to satisfy the type and quantity of medicine.

[0072] The medicine management device according to paragraph 4 makes it possible to reliably order the type and amount of medicine required for a scheduled medical procedure without requiring any action by the user.

[0073] (Section 5) The drug management device according to paragraph 5 is a drug management device according to any one of paragraphs 1 to 4, further comprising: A learning model creation unit that updates the learning model by performing additional machine learning on the learning model using training data including information on the type and amount of medicine used in the medical care-related procedure, and constructs a new classifier using the updated learning model. Equipped with.

[0074] In the medicine management device according to paragraph 5, even if there is a change in medical personnel working at a medical institution, it becomes possible to estimate more accurate medicine ordering information based on the patterns after the change. [Explanation of symbols]

[0075] 1. Drug management system 10. Medical institution system 11…Medical institution server 12...Storage section 121...Patient information storage unit 122...Medical worker information storage unit 123... performance information storage unit 124...Reservation information storage unit 125…Medical Procedure Database 13... Classifier 141…Learning model creation department 142…Schedule entry reception section 143…Order Information Creation Department 144...Display control unit 145…Order Information Editing and Reception Department 146…Order Execution Department 147…Order Data Acquisition Department 148...History data acquisition section 149...Analysis processing section 18...Input section 19...Display section 21…Department terminal 22...Storage section 221...Patient information storage unit 222…Medical worker information storage unit 231…Reservation Information Reception Department 232...Information transmission unit 28...Input section 29...Display section 31... Contractor device

Claims

1. a classifier constructed by a learning model created by machine learning using training data including information on the schedule of past medical-related procedures performed at medical institutions and information on the types and amounts of medicines used in the medical-related procedures, where information on the schedule of the medical-related procedures is input and information on the types and amounts of the medicines is output; a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on information on the type and amount of drug output from the classifier; A medicine management device comprising:

2. The information on the schedule of the past medical-related procedure includes information on the day of the week, The schedule input receiving unit receives input of schedule information including information on days of the week. The drug management device according to claim 1 .

3. moreover, A display unit; a display control unit that displays the order information on the display unit; an order information edit receiving unit that receives edits of information on the type and amount of medicine included in the order information; The drug management device according to claim 1 or 2, further comprising:

4. moreover, a medical care-related procedure database that stores information on the types and amounts of medicines required when performing the medical care-related procedure; a reservation information storage unit in which reservation information for medical procedures accepted at the medical institution is stored; Equipped with The order information edit receiving unit identifies the type and quantity of medicines required for the reserved medical-related procedure stored in the reservation information storage unit based on the medical-related procedure database, and if the type and quantity of medicines included in the order information are insufficient, modifies the order information so as to satisfy the type and quantity of the medicines. The medicine management device according to claim 3 .

5. moreover, a learning model creation unit that updates the learning model by performing additional machine learning on the learning model using training data including information on the types and amounts of medicines used in the medical care-related procedure, and constructs a new classifier using the updated learning model; The drug management device according to claim 1 or 2, further comprising:

6. a computer in which a classifier constructed by a learning model created by machine learning is stored, the learning model inputting information on the schedule of medical-related procedures performed in the past at the medical institution and outputting information on the type and amount of medicines used in the medical-related procedures, using training data including information on the schedule of the medical-related procedures performed in the past and information on the type and amount of medicines used in the medical-related procedures; a schedule input receiving unit that receives input of information about future schedules; an order information creation unit that inputs the future schedule into the classifier and creates drug order information based on the information on the type and amount of drug output from the classifier; A medicine management program characterized by operating as follows.

7. a storage unit that stores information on the schedule of past medical procedures performed at a medical institution and information on the types and amounts of medicines used in the medical procedures; a learning model creation unit that creates a learning model by machine learning using training data including information on the schedule of past medical-related procedures performed at the medical institution and information on the types and amounts of medicines used in the medical-related procedures, and that inputs information on the schedule of the medical-related procedures and outputs information on the types and amounts of medicines; A learning model creation device comprising:

Images