A program that has a hospital bed management function and supports healthcare professionals in selecting medical resources and patients.

Abstract

When determining which treatments and medical procedures to provide to patients from the medical resources held by the hospital, the hospital aims to optimize its revenue by systematizing the decision-making process for medical resources, while also using computer-based predictions of patient utilization of medical resources, thereby increasing the efficiency and profitability of resource utilization. [Solution] Patient information and information on available medical resources that the medical institution can provide to the patient are stored. Based on the patient information, a prediction of the medical resources needed by the patient is calculated as patient utilization prediction resource information, and the patient utilization prediction resource information is matched with the available medical resource information (S707). The matching results are presented to the medical professional, and input from the medical professional is accepted. The patient utilization prediction resource information is changed to patient utilization planned (decided) resource information (S710), thereby accepting the patient's selection and providing the medical resources to the patient.

Description

Technical Field

[0006] , , ,

[0001] The present invention relates to a program (information processing apparatus and method) used in a hospital bed management system of a medical institution, and particularly supports the determination of medical resources and the determination of patients by medical staff.

Background Art

[0002] The current situation of hospital management in Japan is not good. There are policy-based medical treatments that are not profitable and research-priority hospitals, public hospitals that cannot get out of the red-ink situation, and private hospitals that are difficult to operate because they have to survive under the same conditions as public hospitals even if they have excellent management capabilities.

[0003] To achieve stable and improved management, the most important thing is to improve the profitability rate. For this purpose, while effectively utilizing medical resources, it is necessary to suppress expenses and eliminate waste. Medical resources are diverse, including physical resources such as hospital beds and examination rooms, human resources of medical staff such as doctors, nurses, medical technicians, and clerks, hard resources such as medical equipment (especially high-cost medical equipment) like CT and MRI, and soft resources such as medical materials and drugs.

[0004] In addition, it is also important to provide medical resources suitable for hospital management and optimize the choices of patients who receive them. If these medical resource information and patient information can be systematized and managed, and the medical resources can be efficiently selected and distributed to patients who need them, it will contribute to improving the management efficiency of hospital management.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

[0026]

[0052] )

[0006] Patent Document 1 describes a technology for optimizing the use of medical resources, and attempts to reduce the probability of readmission as a measure to optimize and minimize medical resources. In particular, in order to reduce negative revenue for medical institutions, it reduces the number of readmissions within 30 days through medical intervention. Requirements for medical intervention, such as a first-line predictive medical institution and a second-line predictive medical institution, are defined, and it is determined which treatments should be provided to the patient at a medical institution (including home care).

[0007] In other words, while it is beneficial in that it leads to a reduction in public healthcare costs by suppressing readmissions and by sharing limited medical resources among multiple medical institutions, there is no disclosure of the prediction, planning, and decision-making process, nor is there any disclosure regarding the optimization of revenue and the use of medical resources at a particular medical institution, or even from the perspective of patient choice.

[0008] Healthcare professionals such as doctors and nurses are required to manage the availability and remaining resources of medical institutions while selecting and providing the necessary resources to patients. In addition, while it is of paramount importance to provide existing medical resources to patients who need them, there is no system in place to support medical institutions in determining which patients should receive medical resources from a business perspective. [Overview of the project] [Problems that the invention aims to solve]

[0009] This invention has been made in view of the above problems, and aims to provide a hospital bed management tool (program, information processing device (system for linking devices), method) for optimizing hospital revenue by systematizing the decision-making of medical resources based on management, while emphasizing computer-based prediction of patient utilization of medical resources when deciding on the provision of treatments and medical procedures to patients from the medical resources held by the hospital, thereby improving the efficiency and profitability of resource utilization. [Means for solving the problem]

[0010] To solve the above problems, the program of the present invention is based on the following principles.

[0011] (1) A program for operating a computer comprising a processor and memory, wherein the memory stores a database, the database stores patient information and information on available medical resources that a medical institution can provide to a patient, and the program has the following functions: a function to calculate a prediction of the medical resources needed by a patient as patient utilization prediction resource information based on the input patient information; a function to match the patient utilization prediction resource information with the available medical resource information and present it to a medical professional; and a function to accept input from a medical professional and change one or more of the patient utilization prediction resource information to either patient utilization plan resource information that plans the medical resources needed by the patient or patient utilization decision resource information that determines the medical resources needed by the patient. Furthermore, as one aspect of this program, after determining the patient's planned resource information or the resource information the patient has decided to use, the program includes updating the information with the remaining available medical resources and continuing the following matching process.

[0012] According to the present invention, patient information and information on available medical resources that can be provided to the patient are stored. Alternatively, patient information that is updated as needed is also stored. Based on the patient's medical condition and disease, etc., included in the patient information, the computer predicts the medical resources needed and calculates patient utilization prediction resource information. That is, the computer first predicts the medical resources that the patient needs. Secondly, the computer matches the patient utilization prediction resource information with the available medical resource information and attempts to determine whether the resource information matches, does not match, or is similar. If there is resource information that matches or is similar and suitable, the medical professional who refers to the matching results makes a selection and changes one or more of the patient utilization prediction resource information to patient utilization planned resource information or patient utilization decided resource information. That is, "predicted" is changed to "planned" or "decided," and the provision of medical resources to the patient is planned or decided. Alternatively, if the status of "predicted" is not changed, the medical resources will not be provided to the patient. This helps to support the operation of hospital bed management tools from a management perspective by presenting to doctors and other personnel which patients the computer has predicted should be allocated to from a management and management efficiency standpoint. Furthermore, it is possible to shift this perspective to that of community healthcare. In one form of this program, the above medical resource information can be linked and recorded with clinical outcomes recorded in electronic medical records, patient-reported outcomes (PROs), QALYs, etc., provided through patient questionnaires, and a formula can be created to determine the most effective use of limited resources in the community and hospital from a clinical and outcome perspective.

[0013] The database contains hospital bed data, facility standards information, etc., and the program has the function of calculating one or more necessary medical resources and / or patient-specific management indicators from the information of one or more patients, using DPC information, hospital bed data, facility standards, patient data, inpatient candidate information, and medical resource information.

[0014] (2) The database stores past DPC information or claims information of medical institutions, and the patient utilization prediction resource information is calculated based on the patient's past DPC information or claims information.

[0015] According to the present invention, in addition to (1), the database stores past DPC information or claims information of medical institutions, and the patient utilization prediction resource information is calculated based on the patient's past DPC information or claims information. Therefore, it is possible to calculate the patient utilization prediction resource information based on information such as the patient's treatment history and appropriate payment amount obtained from past DPC information or claims information. For example, it is useful for predicting medical resources such as whether to perform an examination using an endoscope or a PET scan.

[0016] (3) The database stores electronic medical record information, and the program has the function of extracting and outputting information into DPC information based on the electronic medical record information, and the DPC information includes consent or non-consent information regarding the secondary use of personal information output from the electronic medical record information.

[0017] According to the present invention, by collecting DPC information or claims information from multiple hospitals and gathering the aforementioned information, it is possible to optimize the use of hospital beds as part of a regional medical plan. On the other hand, collecting all claims information may create resistance from each hospital, so the invention also includes implementing a method to enable provision by automatically masking, modifying, or omitting some of the information. Note that DPC information and claims information are payment information submitted to the payment fund once a month, so the data is updated about once a month. On the other hand, electronic medical record data is recorded each time a medical professional examines a patient, so the data is updated more than once a day during hospitalization. Furthermore, image data and the data used for image analysis of those images are usually updated in real time. This research utilizes the fact that DPC information and claims information, which are updated and finalized about once a month, are easy to output and easy to anonymize, and links them to more real-time information for management and analysis. Furthermore, there are differences in the quality and structure of this information. DPC data, being payment information, is strictly standardized in terminology, structured, and highly accurate. On the other hand, electronic medical record information is poorly structured. Real-time information may be of lower quality. While all of this information is evaluated comprehensively, linking other information to the highest quality DPC information, which is the most structured and has been evaluated by many people, can improve the accuracy of predictions.

[0018] (4) The changes are made by changing the predictions to plans, thereby accepting the patient's choice.

[0019] According to the present invention, when a healthcare professional changes "prediction" to "scheduled" or "decided," the computer accepts the patient's selection, which in turn schedules patient resource information to be provided to the patient, and the patient's hospital bed admission is decided. According to this invention, by recording logs of cases where medical professionals did not manually change the schedule based on computer-generated predictions, along with the reasons for doing so, it is possible to identify challenges and solutions when machine learning-based predictions are not adopted, thereby contributing to improving the performance of machine learning. Alternatively, by visualizing the reasons why medical professionals rejected AI predictions, it is possible to manage medical professionals. Furthermore, by implementing procedures such as requiring approval from a supervisor before medical professionals can change matching results, it is possible to increase compliance with operational guidelines. It should be noted that machine learning can be used not only for prediction but also for matching, and the recorded reason information mentioned above may be used to improve the performance of matching.

[0020] (5) If the forecast is not changed to a plan when the above change is made, a log and / or the reason shall be recorded.

[0021] According to the present invention, by recording the log and the reason when a medical worker does not manually change the computer-based prediction schedule, it is useful for deriving problems and solutions when the AI prediction is not adopted, and contributes to improving the performance of machine learning by AI. Alternatively, visualizing the reason why a medical worker does not adopt the AI prediction is useful for managing the medical worker.

[0022] (6) The medical resource information includes single-function resources and multi-function resources, and the multi-function resources are defined as a set of replaceable single-function resources. In one example, as single functions, simple hospital beds (beds used only for specific patient care or treatment) and nurses with practical experience in dealing with one department can be cited. As examples of multi-function resources, mixed hospital beds (beds capable of dealing with patients in multiple departments) and nurses who can rotate through multiple departments to deal with various patients can be cited. Doctors who can be seen in both mixed wards and doctors who can perform multiple surgeries are judged as multi-function resources.

[0023] According to the present invention, the medical resource information includes simple hospital beds, mixed hospital beds, operating rooms, skilled nurses, and also includes an increase in resources by doctor office work assistance and DX tools. By comprehensively classifying them into single-function resources and multi-function resources and defining the multi-function resources as a set of replaceable single-function resources, the complicated resource allocation can be simplified, and the resource management can be systematically and efficiently performed. Although doctor resources are precious, by arranging doctor office work assistance or purchasing DX tools, the work efficiency of doctors can be improved, and as a result, the available resources of medical institutions can be increased. In one aspect of this program, the increased part of the resources is also calculated and presented, enabling the available resources to be increased with assistance and accurately estimated.

[0024] (7) When making the decision, it is made by referring to the selection support information obtained by associating the patient utilization prediction resource information and the available medical resource information. (8) The selection support information is at least one of the sales amount calculated from any one or more of the receipt information, DPC information, and facility standard information stored in the database, the expenditure required for providing the patient-required medical resources, and the profit obtained by dividing the sales amount by the expenditure. (9) When making the decision, any one or more of the prediction information of the sales amount prediction, the expenditure prediction, and the profit prediction are referred to.

[0025] The selection support information (including management index information) is the calculation result based on the balance data obtained by dividing the revenue including receipts calculated from the data including DPC information and facility standard information stored in the database by the expenditure required for providing the patient-required medical resources necessary for the patient and / or the past balance data based on the past management index information. DPC information refers to information submitted by medical institutions to the Payment Fund, specifically the information recorded in Forms 1, 3, 4, E file, F file, D file, Outpatient E file, and Outpatient F file. Specifically, DPC information includes the following: First, information about each facility, such as facility name and facility code. Second, information about each patient, such as date of birth, gender, height, weight, smoking index, pregnancy information, newborn information, and elderly information. The third item is information for each patient's hospitalization, including the primary illness / injury, the illness / injury that led to hospitalization, the illness / injury that required the most medical resources, ICD-10 code, modifier code, the department that treated the "illness / injury that required the most medical resources," the reason for hospitalization, the purpose and progress of treatment, previous discharge, readmission survey, re-transfer survey, date of admission, payload, date of admission, date of admission, route of admission, whether or not the patient was referred from another hospital, whether or not the patient was admitted from our hospital's outpatient clinic, whether or not the patient was admitted for scheduled or emergency medical care, whether or not the patient was transported by ambulance, whether or not there was any self-injury or suicide attempt, whether or not there was an overdose, date of discharge, destination of discharge, outcome at discharge, and whether or not home medical care was provided after discharge. This includes information such as whether or not there were repeated short-term hospitalizations (chemotherapy, radiotherapy, etc.), whether or not clinical trials were conducted, whether or not the patient was hospitalized in a general ward, a psychiatric ward, or other ward, whether or not there were pressure ulcers, whether or not there were pressure ulcers at the time of admission, the use of medical resources, gestational age at admission, birth weight, gestational age at birth, diagnostic information (comorbidities, sequelae, intractable diseases, etc.) for the disease that required the most medical resources, anesthesia (intravenous anesthesia, spinal anesthesia, etc.), initial or recurrence of cancer, TNM classification of cancer, whether or not chemotherapy for cancer was administered, severity of pneumonia, angina pectoris, information on patients with chronic ischemic heart disease, modified Rankin Scale at discharge, onset time of heart failure, systolic blood pressure, heart rate, cardiac rhythm, dementia-specific group homes, elderly day service centers, whether or not the disease that required the most medical resources was cured or improved, whether or not there was remission, whether or not there was no change, whether or not there was exacerbation, information on death or death by other means, information on medications used, prescriptions, and tests performed. In particular, the drug information, prescription details, and test details recorded in the EF file include information on the actual drug resources and test resources used, making it effective for predicting drug resources and test resources.Furthermore, regarding drug resource information, by adding information not included in the DPC information shown below, such as the prescribing physician, the prescribing physician's supervising physician, and the patient's primary physician, it is possible to clarify the decision-makers for administering drug resources.

[0026] This invention streamlines hospital resource management by adding the following information to each patient's DPC information: specific ward used, specific nurse in charge, patient information of that nurse, attending physician, patient information of the attending physician, clinical pathway information of the ward, information described in the electronic medical record template, progress notes in the electronic medical record, summary information in the electronic medical record, content described in the profile section of the electronic medical record, information from the electronic questionnaire, consent / non-consent information regarding personal information, bed information of each medical institution, physician who prescribed medication, supervising physician of the physician who prescribed medication, patient's primary physician, resource information, image information, information converted from image information to text, hospital employee survey results, employee health check data, and information on employee chat communication frequency and sentiment analysis. In particular, the information described in the admission template by the admission / discharge support center in the electronic medical record template is useful for predicting the length of hospital stay. In addition, information related to medical fee additions is also included and is useful for predicting management indicators. Information from electronic questionnaires entered by patients is also useful. Typically, asking patients to fill out an electronic questionnaire once a day during their hospital stay allows for structured collection of information regarding their discharge wishes and changes in their physical condition, improving predictive capabilities. Furthermore, it becomes possible to collect employee survey data, including employee motivation, work performance, dissatisfaction, willingness to implement work-style reforms, and understanding of acceptance testing, which is effective for human resource management. Additionally, by introducing employee chat, it is possible to evaluate employee motivation and use information from the frequency of communication and sentiment analysis of conversations to manage resources. Based on this information, process mining can be used to analyze the hospital's workflow and resource usage in detail, enabling the identification of continuous improvement measures to enhance operational efficiency and the quality of patient care. By using process mining technology, it is possible to visualize work procedures and resource consumption patterns in each department of the hospital, identifying unnecessary processes and excess or insufficient resources.Based on this analysis, it is possible to support data-driven decision-making to optimize operations and effectively utilize resources, leading to measures that improve the quality of patient care and reduce the burden on staff. Furthermore, by using a waterfall chart, actual performance against targets for ward occupancy rates and operating room occupancy rates can be visually displayed, and resource usage can be monitored and analyzed in real time. Waterfall charts are suitable for clearly representing which elements account for the largest proportion of a given composition, and which elements have a significant impact on changes over time. Using this chart, it is possible to quickly identify the causes of decreased occupancy rates or resource shortages during a specific period and take countermeasures. For example, if ward occupancy rates are low on a particular day or time, the cause can be analyzed, and data-driven optimal improvement measures can be implemented, such as adjusting staff allocation or patient admission and discharge schedules. Furthermore, based on patient information, it is possible to search for similar patients from the past and use the medical fee claim information and DPC information from those patients as reference information to assign medical fee claim information, DPC information, and management indicators to the currently viewed patient. In this process, it is also possible to specify how broad the similarity range should be, present patients with good management indicators from the information of similar patients, and use that information as a reference to assign medical fee claim information and DPC information to the currently viewed patient.

[0027] (10) The prediction of medical resources is made by using one or more of the following information stored in the database: past claims information, DPC information, facility standards information, clinical pathway information, bed utilization information, assigned nurse information, and surgery time, or one or more of the following information: the patient's diagnosis, surgery name, DPC information, claims information, clinical pathway, and electronic medical record template information, to predict one or more of the following: length of stay in the hospital bed, type of hospital bed, surgery time, required nursing time, required nursing skills, sales, expenses, and profits. (11) The program makes the prediction using one or more of the patient's past claims information, DPC information, facility standards information, and clinical pathway information stored in the database. (12) The program uses one or more of the patient's past claims information, DPC information, facility standards information, and electronic medical record template information stored in the database to make predictions for one or more of the following: sales, expenses, and profits.

[0028] (13) For the calculation of the above prediction, the past number of days in hospital bed, type of hospital bed, operating time, primary physician, primary nursing skills, sales, expenses, and profits of patients who match either the disease name or the disease name to which the most resources were invested are aggregated and statistics are calculated. "Statistics" are numerical indicators that represent the characteristics of the data and include the mean, median, mode, quantiles, variance, and standard deviation. These are used to understand the central tendency, variability, and distribution of the data. (14) For the calculation of the above prediction, a similarity search is used to obtain information on past patients, and sales, expenses, and profits are aggregated to calculate statistics. Similarity search is the act or process of finding necessary items from data or materials in order to find specific information. Types include full-text search, logical search, vector search, semantic search, matching search, and regular expression search, and it is possible to efficiently obtain information using different algorithms and technologies for each. As a further form of search, when using progress notes and summary information of electronic medical records, multiple pathologies and multiple symptoms are usually described in one article, but it is possible to use generation AI to separate and describe each pathology and the findings / examination findings related to that pathology. Furthermore, examination findings can be described by examination name, adjective, finding, and location (example notation: CT=2cm*hemorrhage@brain). Using a large-scale multimodal model (LMM), it is possible to generate text by dividing image data into examination name, adjective, findings, and body part. By searching this output as a single unit, it is possible to find cases with similar examination findings. Furthermore, by searching for the underlying pathological conditions in these cases with similar examination findings based on the information described above, it is possible to implement a system that is effective in assigning DPC (Diagnosis Procedure Combination) diagnoses.

[0029] (15) A machine learning model is trained for the calculation of the above prediction, and the trained predictive machine learning model is used. According to the present invention, vector search maps data to a high-dimensional vector space and allows information to be retrieved based on similarity, requiring multiple searches. In contrast, semantic search allows the search engine to understand the user's intent and the meaning of the query, not just perform simple keyword matching searches, and provide highly relevant information. For example, it is possible to extract and search only for abnormal findings. As reference information, it is also possible to search for similar patients from the past based on patient information, and use the medical fee claim information and DPC information for those patients as reference information to assign medical fee claim information, DPC information, and management indicators to the patient currently being viewed. In this case, it is possible to specify how broad the similarity range should be, to obtain highly specific reference information, or to obtain highly sensitive reference information and select patients with good management indicators from among them, and then assign medical fee claim information and DPC information to the patient currently being viewed based on the information of those patients.

[0030] (16) When calculating the available medical resource information, the program performs image analysis on image information taken within one or more facilities such as operating rooms, ICUs, and wards, converts it into category information, and works in conjunction with a function that evaluates the utilization of hospital beds or the level of staff busyness. (17) When calculating the available medical resource information, the program is linked to a function that performs image analysis (OCR processing) on ​​image information captured from one or more of the operating room monitor information, ICU monitor information, or ward monitor information to evaluate the patient's vital signs. Image analysis includes classification, representation extraction using large multimodal models, clustering, time series analysis, object detection, person tracking, image classification, segmentation, and OCR (optical character recognition). By combining these methods, it becomes possible to determine whether beds are available, the progress of surgery, whether patients are in operating rooms, and how many nurses are assigned to operating rooms and wards, thereby enabling a multifaceted evaluation of the utilization of medical resources and the efficiency of staffing.

[0031] According to the present invention, real-time information can be obtained through image analysis, and by evaluating the occupancy rate of hospital beds, the level of staff busyness, and the vital signs of patients, it contributes to the increased efficiency of hospital bed management.

[0032] As one manifestation of the aforementioned program, when calculating the available medical resource information, it has a function to extract and output profile information / template information from electronic medical records as DPC information, and also a function to periodically extract, output, and save the DPC information with an output time assigned to it. As a result, it becomes possible to treat how an individual's DPC information has been input, modified, and changed over time as input for machine learning. Consequently, machine learning can learn the management and operation of DPC information in that hospital, and it becomes possible to improve prediction performance. In fact, it is possible to do this and improve prediction performance. Another manifestation is that it is possible to include an anonymized hash ID in the DPC information by hashing the patient ID as a seed. Similarly, an anonymized hash ID is assigned to information not listed in the DPC information by hashing the patient ID as a seed using the same method, and it becomes possible to link the DPC information with the information not listed in the DPC information based on this anonymized hash ID. In fact, it is possible to do this and reduce the effort required to link information from which the patient ID has been removed. Another manifestation is that it is possible to feed the output of DPC information into machine learning with group information attached. Specifically, the drug information included in the EF file contains drugs with various dosages and administration methods. However, it is possible to assign group information to these products, such as "intravenous drugs," "hypertension drugs," or "amlodipine (including multiple administration methods)," and then perform machine learning on this group information. In fact, doing so can improve predictive performance.

[0033] According to the present invention, DPC information and claims information are payment information submitted to the payment fund once a month, and are therefore typically completed about once a month. On the other hand, electronic medical record data is recorded each time a medical professional examines a patient, and is therefore typically updated at least once a day during hospitalization. Furthermore, image data and the data used for image analysis of those images are typically updated in real time. The use of these information sources with different output frequencies is also a feature of the present invention. In addition, since DPC information and claims information are deduplication and standardization, output, anonymization, and understanding of their operation at each hospital are easy. Because the information is standardized, it is possible to reduce the implementation costs at each hospital. Taking advantage of this, electronic medical record information and image analysis information are linked to DPC information, enabling predictions using real-time situations. Furthermore, real-time accuracy is crucial for predictions; if a patient has already been discharged by the time a prediction is made, the prediction becomes meaningless. To address this, one aspect of the present invention makes it possible to output and analyze DPC information periodically with added output time, and simultaneously link it with other highly real-time information, making it possible to observe changes over time. By combining this function with logs of responses to healthcare professional matching and records of reasons for changes, it becomes possible to conduct efficient review meetings. This is an effective cause analysis tool (waterfall) and process mining that improves the decision-making abilities of healthcare professionals, and AI makes improvements based on feedback.

[0034] (19) The program has a function to aggregate and output either DPC information or claims information from multiple hospitals based on the search (query) information when calculating the available medical resource information, and a function to not display aggregated values ​​below a threshold as specific numerical values ​​when aggregating and displaying the aggregated values.

[0035] According to the present invention, by collecting DPC information or claims information from multiple hospitals and aggregating this information, it can be used to optimize hospital bed utilization as part of a regional medical plan. Furthermore, since collecting all claims information may create resistance from individual hospitals, the invention also includes implementing a method to enable provision by automatically masking, modifying, or omitting some information, i.e., by having a hiding function.

[0036] (20) The program causes the processor to perform the following steps: a first step of replacing a multifunctional resource with one or more substitutable single-function resources to create a single-function resource dataset; a second step of predicting the medical resources required for patients requiring hospitalization from the patient information; a third step of calculating the required medical resources for each candidate patient from among the candidates for hospitalization; and a fourth step of identifying patients requiring hospitalization from the single-function resource dataset, the medical resources required for hospitalization, and the required medical resources. Furthermore, the program causes the processor to perform the following steps: a fifth step of calculating patient-specific management indicators (selection support information) for the patients requiring hospitalization; and a sixth step of displaying the patient-specific management indicators (selection support information) and accepting patient selection.

[0037] According to the present invention, the first step (creation of a single-function resource dataset) is effective in building a system that can handle multiple diseases. Step 2 (predicting the medical resources needed for patients admitted) predicts the medical resources needed for each candidate patient admitted, and Step 3 (calculating the medical resources needed for each candidate patient admitted) calculates the medical resources needed for each candidate patient admitted.

[0038] Step 4 (Identification of Admitted Patients) involves matching various information from the single-function resource dataset obtained in Step 1, the medical resources required by admitted patients obtained in Step 2, and the medical resources required by Step 3 to identify admitted patients, thereby presenting admitted patients who are compatible with the remaining resources of the medical institution. In addition, in step 5 (calculation of patient-specific management indicators), management indicators are calculated for the patients identified in step 4, and linked to these indicators on a patient-by-patient basis. The relationship between patients and management indicators is clearly presented, which helps doctors and other medical professionals in selecting patients for hospitalization in step 6, and assists them in making decisions based on computer-generated resource predictions. [Effects of the Invention]

[0039] The program (information processing device and method) of the present invention provides a hospital bed management system for optimizing the revenue of medical institutions and effectively supports medical professionals in selecting medical resources and patients. [Brief explanation of the drawing]

[0040] [Figure 1] A diagram showing the overall configuration of a system according to an embodiment of the present invention. [Figure 2] A block diagram showing an example of the functional configuration of a server. [Figure 3] A diagram illustrating the overview of the hospital bed management system according to this embodiment. [Figure 4] A diagram showing the patient allocation situation for each hospital bed. [Figure 5] A diagram illustrating the relationship between medical resources and patients. [Figure 6] A diagram showing the relationship between patient data and management indicators and other information. [Figure 7] A flowchart illustrating the processing procedure of the hospital bed management system according to this embodiment. [Best Mode for Carrying Out the Invention]

[0041] <Overall System Configuration> Figure 1 shows the overall configuration of System 1 according to an embodiment of the present invention. In this System 1, a terminal 10 on the hospital side and an external server 20 are connected via a wired / wireless network 80 such as the Internet. Server 20 is a server that functions as a web server (including a cloud server) and exchanges information with terminal 10 via web pages. In addition, while a web browser for viewing web pages is installed on terminal 10, a dedicated application for enjoying the services of server 20 may also be installed, and the system may be configured so that viewing web pages is possible through the dedicated application.

[0042] Terminal 10 is a device operated by medical personnel (doctors, nurses, laboratory technicians, office staff, assistants, etc.) at a medical institution, and is connected to the server 20 via the network 80. Terminal 10 connects to the network 80 by communicating with communication devices such as a wireless base station 81 that supports various communication standards such as LTE, and a wireless LAN router that supports IEEE and wireless LAN standards. Terminal 10 comprises a communication IF 12, an input device 13, an output device 14, a memory 15, a storage unit 16, and a processor 19. Terminal 10 can be a desktop or laptop PC, a tablet, a smartphone, or other portable device.

[0043] The communication interface 12 is an interface for the terminal 10 to communicate with external devices and input / output signals. The input device 13 is an input device (such as a keyboard, touch panel, touchpad, mouse, or other pointing device) for receiving input operations from the user. The output device 14 is an output device (such as a display or speaker) for presenting information to the user. The memory 15 is for temporarily storing programs, data processed by programs, etc., and is a volatile memory such as DRAM. The storage unit 16 is a storage device for saving data, such as flash memory or an HDD. The processor 19 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0044] Server 20 is managed by the administrator of System 1 according to an embodiment of the present invention, and stored contents such as modification / addition / deletion of information can be modified as appropriate by healthcare professionals who are users of Terminal 10. Server 20 may also have the functionality of an electronic medical record device, allowing healthcare professionals in a medical facility to view input items and contents of the electronic medical record via a terminal device (not shown) and modify / add to the input contents. Furthermore, the server accepts editing operations of electronic medical record templates and electronic questionnaires performed by healthcare professionals via this terminal device, and the stored contents are modified / added / deleted based on these editing operations.

[0045] Server 20 is a computer connected to network 80 and includes a communication interface 22, an input / output interface 23, memory 25, storage 26, and a processor 29.

[0046] The communication interface is an interface for inputting and outputting signals so that the server 20 can communicate with external devices. The input / output interface 23 functions as an interface to an input device for receiving input operations from the user and an output device for presenting information to the user. The memory 25 is for temporarily storing programs and data processed by programs, etc., and is a volatile memory such as DRAM. The storage 26 is a storage device for saving data, such as flash memory or an HDD. The processor 29 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0047] <Functional configuration of Server 20> Figure 2 is a block diagram showing an example of the functional configuration of server 20. Server 20 comprises a communication means 220, an input device 230, an output device 240, an audio processing unit 270 to which a microphone 271 and a speaker 272 are connected, a storage means 280, and a control means 290, with each block being electrically connected by a bus or the like.

[0048] The communication means 220 performs modulation and demodulation processing for the server 20 to communicate with other devices, processes the signal calculated by the control means 290 for transmission, and transmits it to the external device. The communication means 220 processes the signal received from the outside and outputs it to the control means 290. In this way, the communication means 220 interprets commands or input content and provides them to each means, and also functions as an interface that interprets various display commands issued from the storage means 280 and performs output control.

[0049] The input device 230 is a device used by a user operating the server 20 to input instructions or information, and may be a keyboard, mouse, reader, or touch-sensitive device. The input device 230 also converts the instructions input by the user into electrical signals and outputs the electrical signals to the control means 290. The input device 230 also includes a receiving port that accepts electrical signals input from an external input device.

[0050] The output device 240 is a display device 241 such as an LCD or organic EL for presenting information to the user operating the server 20. The display 241 displays data according to the control content of the control means 290.

[0051] The audio processing unit 270 performs digital-to-analog conversion processing of the audio signal. It converts the signal provided from the microphone 271 into a digital signal and provides the converted signal to the control means 290. The audio processing unit 270 also provides the audio signal to the speaker 272. The audio processing unit 270 is implemented by an audio processing processor, and the microphone 271 receives an audio input and provides the audio signal corresponding to the audio input to the audio processing unit 270. The speaker 272 converts the audio signal provided by the audio processing unit 270 into audio and outputs the audio to an external device connected to the server 20.

[0052] The storage means 280 is implemented by memory (RAM) 25 and storage 26 such as a disk device (floppy disk, hard disk, or magneto-optical disk, etc.) and stores data, programs, etc. used by the server 20. In addition to the application programs of this system, the storage means 280 stores data such as facility standards, electronic medical records, DPC / receipts, hospital beds, patients, available medical resources (single-function resources, multi-function resources), predicted acceptance / rejection logs / reasons, and clinical pathways.

[0053] The control means 290 is realized when the processor 29 reads the application program stored in the storage means 280 and executes the instructions contained in the application program. The control means 290 also controls the operation of the server 20 and the terminal 10 and operates according to the application program, thereby performing functions as an input operation reception unit, a transmission / reception unit, a data processing unit, a display control unit, a matching processing unit, a multi-function resource single-function allocation processing unit, a patient utilization prediction / required resource processing unit, a medical resource remaining amount calculation processing unit, and a selection support / management indicator calculation processing unit.

[0054] The input operation receiving unit processes instructions or information input from the input device 230. Specifically, it receives information based on instructions input from a keyboard, mouse, etc. The input operation receiving unit also receives voice instructions input from the microphone 271. Specifically, for example, it receives voice signals input from the microphone 271 and converted into digital signals by the voice processing unit 270. It also obtains instructions from the user by analyzing the received voice signals and extracting predetermined nouns.

[0055] The transmitting / receiving unit performs processing to enable the server 20 to send and receive data according to the communication protocol. Specifically, for example, the transmitting / receiving unit sends the content entered by the user to the server 20 and receives information about the user from the server 20.

[0056] The data processing unit performs calculations on the data received by the server 20 according to the application program and outputs the calculation results to memory 25 or the like.

[0057] The presentation control unit controls the output device 240 to present information provided by the server 20 to the user. Specifically, for example, the presentation control unit displays the information transmitted from the server 20 on the display 241. The presentation control unit also outputs the information transmitted from the server 20 through the speaker 272.

[0058] <Details of the hospital bed management system> Figure 3 is a diagram illustrating the overview of the hospital bed management system according to this embodiment. (A) manages hospital beds for ICU patients / HCU patients, and (B) manages hospital beds for gastrointestinal patients / cardiovascular patients. The list of potential patients for admission includes a forecast of each patient's business indicator (profit = sales - costs) as selection support information.

[0059] In other words, in the example in Figure 3(A), assuming that the ICU / HCU can accommodate 20 patients, and based on available medical resource information, 10 patients can be admitted to the ICU and 10 to the HCU, with 8 ICU patients following clinical pathways, 8 HCU patients following clinical pathways, and 1 patient scheduled by a physician, the system predicts that 5 ICU patients will be admitted on day 1, 4 on day 2, 2 on day 3, and 1 from day 4 onwards (total number of admissions), while the physician's schedule predicts that 2 ICU patients will be admitted on day 1, 4 on day 2, 1 on day 3, and 1 from day 4 onwards (total number of admissions). The decision of which patients to admit is then made based on the management indicator forecast. In doing so, the admission decision is made based on the list of candidate patients, referring to the patient-specific management indicator forecast for each patient. Furthermore, in the example in Figure 3(B), assuming that there is capacity to admit patients in 6 gastroenterology beds, 6 cardiovascular beds, and 5 mixed beds, and that based on the available medical resource information, 10 gastroenterology patients and 10 cardiovascular patients can be admitted, with 8 gastroenterology patients following clinical pathways and 8 cardiovascular patients following clinical pathways, and 1 patient scheduled by a physician, the system forecast predicts that 5 gastroenterology patients will be admitted on day 1, 4 on day 2, 2 on day 3, and 1 from day 4 onwards (total number of admissions), while the physician's schedule predicts that 2 patients will be admitted on day 1, 4 on day 2, 1 on day 3, and 1 from day 4 onwards (total number of admissions). The decision of which patient to admit to which bed is then made based on the forecast of management indicators. Furthermore, while the system forecasts that 3 cardiovascular patients will be admitted on day 1, 4 on day 2, 2 on day 3, and 1 from day 4 onwards (in total number of admissions), the physician's plan is to admit 2 on day 1, 4 on day 2, 2 on day 3, and 1 from day 4 onwards (in total number of admissions). The decision of which patients to admit to which beds is then made based on the management indicator forecast. In doing so, the admission decision is made based on a list of candidate patients, referring to the patient-specific management indicator forecast for each patient. For example, if gastroenterology patients 1, 2, and 3 and cardiovascular patients 1 and 2 are prioritized, cardiovascular patient 3, whose management indicator forecast is poor, will be excluded from the system as a patient to be admitted, and efforts will be made to collaborate with other medical institutions. Furthermore, Figure 3 shows an example of displaying management indicators, where projected profit, projected sales, costs, and floor space forecast are recorded in a table, and at least one piece of information can be displayed in a list.

[0060] Figure 4 shows a patient allocation table for each hospital bed, illustrating the calculations for determining which bed to allocate patients (P01-P15) to in order to use the beds efficiently. Here, the daily forecast is calculated by a computer based on DPC information, etc., to determine the length of stay for each patient and allocate beds accordingly. The daily schedule is determined by a physician based on the computer's forecast. The collaboration between the computer's forecast and the physician's schedule shows the bed occupancy status for the following day and beyond. In this case, if a bed vacancy is anticipated, the system determines the bed to which the patient will be admitted by selecting the patient with the best management indicators from among the prospective inpatients. In the case of clinical pathways, schedules for multiple days in advance are determined.

[0061] In the example in Figure 4(A), patient P14 is assigned to the hospital bed designated as "Hospitalized 01" on days when patient P01 is available (dashed block), patient P15 is assigned to the hospital bed designated as "Hospitalized 02" on days when patient P02 is available (dashed block), patient P12 is assigned to the hospital bed designated as "Hospitalized 05" after patient P05 (dashed block), and the hospital bed designated as "Hospitalized 06" is left unassigned on days when patient P06 is available (dashed block). The assignment of each patient is determined by extracting data from the list shown in Figure 4(B) (a combined list of patient data, hospitalization forecasts, and management indicators).

[0062] Figures 5 and 6 show the data structure underlying the list shown in Figure 4(B). Figure 5 shows the relationship between medical resources (disease, available beds, and usage / unused status) and patients, while Figure 6 shows the relationship between patient data (patient name, gender, age, occupation, expected bed, bed preference, preferred timing, etc.) and management indicators (sales, costs, profit, profit margin) and other information (priority, category, availability of weekend admissions). Input content for input items can be directly entered by medical professionals using input devices 13, 230, or data entered using input device 13 can be acquired by server 20 via network 80. Note that the black-colored cells in Figures 5 and 6 highlight patients P14, P15, and P12 as potential patients scheduled for admission before hospitalization, and these patients are placed in the available resources in Figure 4(A).

[0063] As shown in Figure 5, for patient P01, the predicted ○ / scheduled ○ (dashed line block B1) indicates that the computer predicts the patient will need to remain in bed for a certain period, and this has already been decided as a scheduled stay by the physician, meaning that continued hospitalization is to secure medical resources. There is a possibility that the physician may later cancel the resource and change the scheduled stay from ○ to -, but normally these resources will be used for patient P01. For patient P09, the prediction is ○ / planned - (dashed line block B2) indicates that while the computer predicts the patient will need to remain in bed for a certain period, it is not yet decided that the patient will continue to stay in bed. After obtaining approval from the physician, it may be decided that the patient will remain in bed, and the planned status may be changed from - to ○. If the prediction is ○ but the planned status is -, the log and reason will be recorded in the prediction acceptance / rejection log / reason data 288. For patient P12, the prediction is - / planned is ○ (dashed line block B3), which indicates that while the computer prediction may indicate the patient may already be discharged / no longer needing hospitalization, the hospital's clinical pathway anticipates continued hospitalization. In such cases, the log and reason may also be recorded in the prediction acceptance / rejection log / reason data 288.

[0064] In other words, the agreement / disagreement between the computer prediction and the physician's schedule is recorded as not being scheduled if the hospitalization prediction is not approved, based on the physician's judgment as the necessity of hospitalization for each patient. The authority to approve the prediction and change it to a scheduled appointment may belong to a medical professional or a medical office assistant directly under the physician.

[0065] As shown in Figure 6, in addition to a list of patient data and management indicators that are always available and linked, other information (priority, category, and whether or not weekend hospitalization is possible) is also available to a limited extent, and when doctors and other staff decide on appointments, they can rearrange the display order based on this other information.

[0066] Medical resources include physical resources such as bed types (simple beds, mixed beds) and operating rooms / laboratories, as well as human resources such as skilled nurses and doctors. The available medical resources are calculated based on the operational status on the day. Pre-admission beds refer to temporary beds where patients are temporarily admitted from a list of potential patients before formal hospitalization.

[0067] Figure 7 is a flowchart showing the processing procedure of the hospital bed management system according to this embodiment. In this system, the control means 290 processes each processing unit via a server 20 connected to a communication line such as the Internet.

[0068] First, a definition process is performed to define multifunctional medical resources as a collection of single-function medical resources (step SA701). The definition process performed in the multifunctional resource single-function assignment processing unit 296 refers to the resource data stored in single-function resource data 286 and multifunctional resource data 287, and the AI ​​makes the optimal assignment and definition based on past definition examples and cases. The AI's learning model is stored as an example of linking single-function resource data 286 / multifunctional resource data 287, and the AI ​​performs machine learning and updates it using prediction acceptance / rejection logs / reason data 288.

[0069] Next, the set of available single-function resources is calculated (step SA702). The single-function resources obtained as a result of the definition process performed in the multi-function resource single-function allocation processing unit 296 are given a "available" flag "1" if they are available, and an "unavailable" flag "0" if they are not available. Availability is determined at the medical institution by referring to DPC information 283 and hospital bed data 284.

[0070] Next, the selection of available medical resources is performed (step SA703). This involves selecting available medical resources from physical resources, human resources, etc., for single-function resources for which the flag "1" has been set in the multi-function resource single-function allocation processing unit 296. In addition, the medical resource remaining amount calculation processing unit 298 analyzes image information taken within one or more facilities such as operating rooms, ICUs, and wards, converts it into category information, and evaluates the remaining amount and availability of medical resources by assessing the utilization of hospital beds or the busyness of personnel, which also contributes to the selection of available medical resources. Furthermore, by analyzing images and evaluating the patient's vital signs, the AI ​​calculates and predicts the remaining amount of available medical resources in the future, which also contributes to the calculation of available medical resource information in real time. As a result, medical professionals can check the available medical resource information in real time, which leads to the decision of whether or not to accept a patient after the determination of the medical resources required by the patient has been made.

[0071] The processing performed by the multi-functional resource single-function allocation processing unit 296 in steps SA701 to SA703 is executed as Task A of the program, which defines a multi-functional resource as a collection of single-function resources and presents it as a medical resource that can be provided.

[0072] Meanwhile, separate from / in parallel with the processing of Task A, the processing of Task B is executed. First, patient data is input into the patient data 285 of the storage means 280 (step SB701). The patient data is stored as patient data 285 by acquiring input information from the input device 13 of the terminal 10 and patient data stored in the storage unit 16. The patient data includes information such as the patient's name, gender, age, occupation, expected hospital bed, bed preference, and preferred timing, as shown in Figure 6.

[0073] Next, for patients in the input / acquired patient data, the system determines whether a clinical pathway (CP) is necessary (step SB702). If it is determined that a CP is necessary, the clinical pathway is entered (step SB704). If it is determined that a CP is not necessary, the system predicts the medical resources required for the patient and calculates the predicted patient utilization resources (step SB703). The predicted patient utilization resources (step SB703) are calculated in the patient utilization prediction / planning / decision resource processing unit 297, which calculates the medical resources required for the patient by referring to electronic medical record data 282, DPC / receipt data 283, etc. For prediction, the system aggregates past hospital stay days, type of hospital bed, surgery time, primary physician, primary nurse skills, sales, expenses, and profits for patients whose disease name or the disease name that required the most medical resources matches, calculates statistics, and makes predictions. For the prediction calculation, semantic search and vector search are used to perform similarity searches to obtain information on past patients, aggregate sales, expenses, and profits, and calculate statistics.

[0074] CP (Clinical Path) determination (step SB702) is made by a computer prioritizing clinical pathway information based on information from electronic medical record data 282, AI-predicted information based on DPC information, or information pre-registered as clinical pathway data 288, or by a physician or other medical professional making a direct determination.

[0075] Next, the patient's required medical resources are determined (step SB705). This is done in the patient utilization prediction / scheduling / determined resource processing unit 297, which calculates the required patient utilization prediction resources (step SB703) or inputs a technical path (step SB704) while referring to patient data 285 and electronic medical record data 282, and determines the patient's required medical resources. Alternatively, if a CP judgment is not required, the calculated patient utilization prediction resources may be determined as the patient's required medical resources.

[0076] Furthermore, electronic medical record data 282 includes an electronic medical record template, and the input items and content of the electronic medical record template include claims information, DPC information, facility standards information, and clinical pathway information. The profile information of the electronic medical record can be extracted and output as DPC information.

[0077] The processing performed in steps SB701 to SB705, which calculate the medical resources required by the patient, is executed as Task B, which links the patient with the medical resources required by that patient.

[0078] The execution results of both tasks A and B are obtained through processing, and a matching process is performed between them (step S707). The matching is performed by the data processing unit, etc., and the computer determines whether there is a match or mismatch between the medical resources that the medical institution can provide (Task A) and the medical resources required by the patient (Task B) (step S708). If the available medical resources and the medical resources required by the patient match as a result of the determination, the matching is determined to be successful, and a process of linking the medical resources required by the patient with management indicators may be performed. The selection support / management indicator calculation processing unit 299 calculates selection support information (management indicators) when providing medical resources required by the patient.

[0079] If the determination in step S708 finds a mismatch between the available medical resources and the medical resources required by the patient, the matching process (step S707) is performed again to substitute with similar medical resources. For example, if there is a mismatch in the human resources among the available medical resources, but there are alternative human resources available, the data processing unit performs the substitution and matching and determines that a match is possible.

[0080] Next, the selection support / management indicator calculation processing unit 299 updates the selection support information and management indicators in response to the successful matching with the patient's required medical resource information generated in step S705. The physician or other personnel then refer to the results of the matching process and the linking process between the selection support information (management indicators) to change the predicted patient utilization resources to planned patient utilization resources and schedule patient admission (step S710). As a result, the planned patient utilization resources are scheduled (step S710), the patient selection and prediction → schedule change process is completed, and this subroutine terminates.

[0081] The results of linking patient-needed medical resources with selection support information (management indicators) from among the available medical resources are displayed and visualized on an output device 240 such as a display 241. The system waits for the doctor or other medical professional to accept the selection, and once a selection is made, the computer determines the status of the prediction as a plan and confirms the patient's acceptance of the selection (step S711), and this subroutine terminates.

[0082] On the other hand, if acceptance is deemed unacceptable in step S709, the unacceptable status is logged, and the medical professional is prompted to record the reason for the unacceptable status (step S711), after which this subroutine terminates. [Industrial applicability]

[0083] This invention is useful in optimizing the revenue of medical institutions and in supporting healthcare professionals in selecting patients and medical resources from the perspective of improving management efficiency.

[0084] [Note A1] A method performed by a computer having a processor and memory, wherein The aforementioned memory stores the database, The aforementioned database stores patient information and information on available medical resources that a medical institution can provide to the patient. The program includes the step of calculating a prediction of the medical resources needed by the patient as patient utilization prediction resource information based on the input patient information, The steps include matching the patient utilization prediction resource information with the available medical resource information and presenting it to healthcare professionals, A method characterized by comprising the step of receiving input from a healthcare professional and making changes from the patient utilization prediction resource information to one or more of the patient utilization planned resource information or patient utilization decided resource information. [Appendix A2] The aforementioned database stores past DPC information or claims information of medical institutions. The method described in Appendix A1, characterized in that the patient utilization prediction resource information is calculated based on the patient's past DPC information or claims information. [Note A3] The aforementioned database stores electronic medical record information. The aforementioned program has a function to extract and output information into DPC information based on electronic medical record information. The method described in Appendix A1 and A2, characterized in that the DPC information includes consent or non-consent information regarding the secondary use of personal information, which is output from the electronic medical record information. [Note A4] The method for making changes to the aforementioned appendices A1 to A3 is characterized by making changes to the predictions to allow the patient to accept the choice. [Note A5] A method characterized by requiring, when the aforementioned A1 to A3 are changed, to record a log or to ask a medical professional to input the reason and record their response if the forecast is not changed to a plan. [Note A6] The method is characterized in that the available medical resource information in the aforementioned appendices A1 to A3 includes single-function resources and multi-function resources, and the multi-function resources are defined as a set of one or more replaceable single-function resources. [Note A7] A method characterized in that, when making the decisions described in the aforementioned appendices A1 to A3, the decision is made by referring to selection support information obtained by linking the patient utilization prediction resource information and the available medical resource information. [Note A8] The method is characterized in that the selection support information in Appendix A7 above is at least one of the following: sales calculated from one or more of the information stored in the database: claims information, DPC information, and facility standards information; expenses required to provide the patient's necessary medical resources; and profit obtained by dividing sales by expenses. [Note A9] A method characterized in that, when making the decision in Appendix A7, one or more forecast information from among the sales forecast, the expenditure forecast, and the profit forecast is referenced. [Note A10] The predictions of the medical resources in the aforementioned appendices A1 to A3 are as follows: The database contains one or more of the following information: past claims information, DPC information, facility standards information, clinical pathway information, bed utilization information, assigned nurse information, and surgical time. Alternatively, one or more pieces of information from the patient's diagnosis, surgery name, DPC information, claims information, clinical pathway, or electronic medical record template information, A method characterized by using to predict one or more of the following: length of stay in hospital beds, type of hospital bed, surgery time, required nursing time, required nursing skills, sales, expenses, and profit. [Note A11] The method described in Appendix A10 above is characterized by performing the prediction using one or more of the patient's past claims information, DPC information, facility standards information, and clinical pathway information stored in the database. [Note A12] The method described in Appendix A10 is characterized by using one or more of the following information stored in the database: past claims information, DPC information, facility standards information, and electronic medical record template information of patients who have previously used medical resources at the medical institution: to make a prediction of one or more of the following: sales, expenses, and profits. [Note A13] A method for calculating the predictions in Appendix A10 above, which involves aggregating the past number of days in hospital beds, type of hospital bed, surgery time, primary physician, primary nurse skills, sales, expenses, and profits of patients who have previously utilized resources at a medical institution and whose disease name or the disease name for which the most resources were invested matches, and then calculating statistics. [Note A14] A method for calculating the predictions in Appendix A10, characterized by obtaining information on past patients by performing a similarity search on patients who have previously used resources at a medical institution, aggregating sales, expenses, and profits, and calculating statistical amounts. [Note A15] A method characterized by using a machine learning model to train a machine learning model using resource usage information of patients who have used resources at a medical institution in the past, for the calculation of the prediction in Appendix A10, and using the trained predictive machine learning model. [Note A16] When calculating the available medical resource information in the aforementioned appendices A1 to A3, A method characterized by combining image analysis (conversion into categorical information) of image information captured within one or more facilities such as operating rooms, ICUs, and wards, with a function to evaluate the utilization of hospital beds or the level of staff busyness. [Note A17] When calculating the available medical resource information in the aforementioned appendices A1 to A3, A method characterized by performing OCR on image information captured from one or more of the following: operating room monitor information, ICU monitor information, or ward monitor information, and linking it with a function to evaluate the patient's vital signs. [Note A18] When calculating the available medical resource information in the aforementioned appendices A1 to A3, A method characterized by having a function to extract and output profile information from electronic medical records as DPC information, and also having a function to periodically extract and output DPC information with an assigned output time. [Note A19] When calculating the available medical resource information in the aforementioned appendices A1 to A3, Based on the search query information, this function aggregates and outputs either DPC information or claims information from multiple hospitals. A method characterized by having a function that, when aggregating aggregated values, does not display those aggregated values ​​as specific numerical values ​​if they fall below a threshold. [Note A20] The processors A1 to A19 mentioned above, The first step is to replace a multi-functional resource with one or more alternative single-functional resources to create a single-functional resource dataset, The second step is to predict the medical resources required for patients who are candidates for admission, based on the aforementioned patient information. The third step involves calculating the necessary medical resources for each candidate patient from among the aforementioned candidates for admission, A fourth step of identifying a patient admitted to bed from the aforementioned single-function resource dataset, the medical resources required for the admitted patient, and the required medical resources, How to make it happen.

[0085] In the following, "computer" is synonymous with "information processing device," and also refers to a "system" formed by the cooperation of computers. [Note B1] A computer comprising a processor and memory, The aforementioned memory stores the database, The aforementioned database stores patient information and information on available medical resources that a medical institution can provide to the patient. The program has a function to calculate a prediction of the medical resources needed by the patient as patient utilization prediction resource information based on the input patient information, The function matches the patient utilization prediction resource information with the available medical resource information and presents it to healthcare professionals. A computer that executes a program having the function of receiving input from a healthcare professional and changing one or more of the patient utilization prediction resource information to either patient utilization planned resource information or patient utilization decision resource information. [Note B2] The aforementioned database stores past DPC information or claims information of medical institutions. The computer that runs the program described in Appendix B1, which has the function of calculating the patient utilization prediction resource information based on the patient's past DPC information or medical claim information. [Note B3] The aforementioned database stores electronic medical record information. The aforementioned program has a function to extract and output information into DPC information based on electronic medical record information. The computer described in Appendix B1 and B2, characterized in that the DPC information includes consent or non-consent information regarding the secondary use of personal information, which is output from the electronic medical record information. [Note B4] The computer is characterized by accepting the patient's choice by changing the aforementioned appendices B1 to B3 to the plan. [Note B5] A computer characterized by either recording a log or requesting a medical professional to input the reason and recording the response if the forecast is not changed to a plan when the above appendices B1 to B3 are changed. [Note B6] The computer is characterized in that the available medical resource information in appendices B1 to B3 above includes single-function resources and multi-function resources, and the multi-function resources are defined as a set of one or more replaceable single-function resources. [Note B7] A computer characterized in that, in making the decisions described in the above appendices B1 to B3, it refers to selection support information obtained by linking the patient utilization prediction resource information and the available medical resource information. [Note B8] The computer is characterized in that the selection support information in the aforementioned appendix B6 is at least one of the following: sales calculated from one or more of the information stored in the database: claims information, DPC information, and facility standards information; expenses required to provide the patient's necessary medical resources; and profit obtained by dividing sales by expenses. [Note B9] A computer characterized in that, in making the decision described in Appendix B7 above, it refers to one or more forecast information from among the sales forecast, the expenditure forecast, and the profit forecast. [Note B10] The predictions of the medical resources in the aforementioned appendices B1 to B3 are as follows: The database contains one or more of the following information: past claims information, DPC information, facility standards information, clinical pathway information, bed utilization information, assigned nurse information, and surgical time. Alternatively, one or more pieces of information from the patient's diagnosis, surgery name, DPC information, claims information, clinical pathway, or electronic medical record template information, A computer characterized by using the following to predict one or more of the following: length of stay in a hospital bed, type of hospital bed, surgery time, required nursing time, required nursing skills, sales, expenses, and profit. [Note B11] The computer described in Appendix B10 above is characterized by performing the prediction using one or more of the patient's past claims information, DPC information, facility standards information, and clinical pathway information stored in the database. [Note B12] The computer described in Appendix B10 above is characterized by using one or more of the following information stored in the database: patient claims information, DPC information, facility standards information, and electronic medical record template information, to make predictions for one or more of the following: sales, expenses, and profits. [Note B13] A computer that calculates statistics for the predictions in Appendix B10 above by aggregating the past number of days in hospital beds, type of bed, surgery time, primary physician, primary nurse skills, sales, expenses, and profits of patients who have previously used resources at a medical institution and whose disease name or the disease name for which the most resources were invested matches. [Note B14] A computer characterized by performing a search using similarity search on patients who have previously used resources at a medical institution to obtain information on past patients, and then calculating statistics by aggregating sales, expenses, and profits for the calculation of the predictions in Appendix B10. [Note B15] A computer characterized by using a machine learning model to train a machine learning model using resource usage information of patients who have used resources at a medical institution in the past, for the calculation of the prediction in Appendix B10 above, and using the trained predictive machine learning model. [Note B16] When calculating the available medical resource information in the aforementioned appendices B1 to B3, A computer characterized by its ability to analyze image information captured within one or more facilities such as operating rooms, ICUs, and wards (converting it into categorical information) and to work in conjunction with a function that evaluates the utilization of hospital beds or the level of staffing. [Note B17] When calculating the available medical resource information in the aforementioned appendices B1 to B3, A computer characterized by its ability to analyze (optically recognize) image information captured from one or more of the following: operating room monitor information, ICU monitor information, or ward monitor information, and to work in conjunction with a function to evaluate the patient's vital signs. [Note B18] When calculating the available medical resource information in the aforementioned appendices B1 to B3, A computer characterized by having the function to extract and output profile information from electronic medical records as DPC information, and also having the function to periodically extract and output DPC information with assigned output times. [Note B19] When calculating the available medical resource information in the aforementioned appendices B1 to B3, Based on the search query information, this function aggregates and outputs either DPC information or claims information from multiple hospitals. A computer characterized by having a function that, when aggregating aggregated values, does not display those aggregated values ​​as specific numerical values ​​if they fall below a threshold. [Note B20] The processors described in the appendix B1 to B19 above are: The first step is to replace a multi-functional resource with one or more alternative single-functional resources to create a single-functional resource dataset, The second step is to predict the medical resources required for patients who are candidates for admission, based on the aforementioned patient information. The third step involves calculating the necessary medical resources for each candidate patient from among the aforementioned candidates for admission, A fourth step of identifying a patient admitted to bed from the aforementioned single-function resource dataset, the medical resources required for the admitted patient, and the required medical resources, The computer that runs it. [Explanation of symbols]

[0086] 10: Terminal (13: Input device, 14: Output device, 15: Memory, 16: Storage unit, 19: Processor) 20: Server (25: Memory, 26: Storage, 29: Processor) 80: Internet network, 81: Wireless base station

Claims

1. A program for operating a computer that includes a processor and memory, The aforementioned memory stores the database, The aforementioned database stores patient information and information on available medical resources that a medical institution can provide to the patient. The program has a function to calculate a prediction of the medical resources needed by the patient as patient utilization prediction resource information based on the input patient information, The function matches the patient utilization prediction resource information with the available medical resource information and presents it to healthcare professionals. A program characterized by having a function that accepts input from healthcare professionals and makes changes to one or more of the following from the patient utilization prediction resource information: patient utilization planned resource information or patient utilization decided resource information.

2. The aforementioned database stores past DPC information or claims information of medical institutions. The program according to claim 1, characterized in that the patient utilization prediction resource information is calculated based on the patient's past DPC information or medical claim information.

3. The aforementioned database stores electronic medical record information. The aforementioned program has a function to output information to DPC information based on electronic medical record information. The program according to claim 1 or 2, characterized in that the DPC information includes consent or non-consent information regarding the secondary use of personal information, which is output from the electronic medical record information.

4. The program according to claim 1 or 2, characterized in that the change is made by changing the prediction to a schedule, thereby accepting the patient's choice.

5. The program according to claim 1 or 2, characterized in that, when the aforementioned change is made, if the prediction is not changed to a plan, it records a log or requests a medical professional to input the reason and records the response.

6. The program according to claim 1 or 2, characterized in that the available medical resource information includes single-function resources and multi-function resources, and the multi-function resources are defined as a set of one or more replaceable single-function resources.

7. The program according to claim 1 or 2, characterized in that the decision is made by referring to selection support information obtained by linking the patient utilization prediction resource information and the available medical resource information.

8. The program according to claim 7, characterized in that the selection support information is at least one of the following: sales calculated from one or more of the information stored in the database: claims information, DPC information, and facility standards information; expenses required to provide the patient's necessary medical resources; and profit obtained by dividing sales by expenses.

9. The program according to claim 7, characterized in that, in making the aforementioned decision, it refers to one or more forecast information, which include the sales forecast, the expenditure forecast, and the profit forecast.

10. The prediction of the aforementioned medical resources is based on one or more of the following information stored in the database: past claims information, DPC information, facility standards information, clinical pathway information, hospital bed utilization information, assigned nurse information, and surgical time. Alternatively, one or more pieces of information from the patient's diagnosis, surgery name, DPC information, claims information, clinical pathway information, or electronic medical record template information, The program according to claim 1 or 2, characterized in that it uses the following to predict one or more of the following: length of stay in hospital beds, type of hospital bed, surgery time, required nursing time, required nursing skills, sales, expenses, and profit.

11. The program is characterized by performing the prediction using one or more of the patient's past claims information, DPC information, facility standards information, and clinical pathway information stored in the database. As the aforementioned selection support information, clinical pathway information is presented. The system accepts input from healthcare professionals and, based on clinical pathway information, collectively changes one or more of the following: patient utilization prediction resource information, patient utilization planned resource information, or patient utilization decision resource information. The program according to claim 10.

12. The program according to claim 10, characterized in that it uses one or more of the following information stored in the database: claims information, DPC information, facility standards information, and electronic medical record template information, to make predictions for one or more of the following: sales, expenses, and profits.

13. The program according to claim 10, which calculates statistics by aggregating the past number of days in hospital beds, type of hospital bed, surgery time, primary physician, primary nursing skills, sales, expenses, and profits of patients who have previously used resources at a medical institution and whose disease name or the disease name for which the most resources were invested matches the aforementioned prediction calculation.

14. The program according to claim 10, characterized in that, in order to calculate the aforementioned prediction, it performs a search using similarity search on patients who have previously used resources at a medical institution to obtain information on past patients, aggregates sales, expenses and profits, and calculates statistics.

15. The program according to claim 10, characterized in that, for the calculation of the aforementioned prediction, a machine learning model is trained using resource usage information of patients who have used resources at a medical institution in the past, and the trained predictive machine learning model is used.

16. The program, in calculating the available medical resource information, The program according to claim 1 or 2, characterized in that it is linked to a function that analyzes image information taken within one or more facilities such as operating rooms, ICUs, and wards to evaluate the utilization of hospital beds or the level of staffing.

17. The program, in calculating the available medical resource information, The program according to claim 1 or 2, characterized in that it is linked to a function that obtains patient vital information by performing OCR on image information captured from one or more of the following: monitoring information from the operating room, monitoring information from the ICU, or monitoring information from the ward.

18. The program, in calculating the available medical resource information, The program according to claim 1 or 2, characterized in that it has a function to extract profile information from an electronic medical record and output it as DPC information, and a function to output the DPC information periodically with an output time assigned to it.

19. The program, in calculating the available medical resource information, Based on the search query information, this function aggregates and outputs either DPC information or claims information from multiple hospitals. The program according to claim 1 or 2, characterized in that it has a function to not display aggregate values ​​below a threshold as specific numerical values ​​during aggregation.

20. The program according to any one of claims 1 to 3 is provided to the processor, The first step is to replace a multi-functional resource with one or more alternative single-functional resources to create a single-functional resource dataset, The second step is to predict the medical resources required for patients who are candidates for admission, based on the aforementioned patient information. The third step is to calculate the necessary medical resources for each candidate patient from among the aforementioned candidates for admission, A fourth step of identifying a patient admitted to bed from the aforementioned single-function resource dataset, the medical resources required by the admitted patient, and the necessary medical resources, A program that executes something.

Images