Electronic health record system

Abstract

[Problem] To provide a system that uses a large language model which has learned a large amount of the latest medical knowledge to assist healthcare professionals create order documents and record documents for electronic health records and thus assists in efficient healthcare activities. [Solution] This electronic health record system that uses a large language model is provided with an electronic health record item extracting means for extracting, from an electronic health record, information on an item necessary for processing a document of any designated document category, and a means for transcribing electronic health record content to a prompt, the means providing the extracted electronic health record content as a prompt for the large language model.

Description

Electronic Medical Record System

[0001] The present invention relates to an electronic medical record system that uses a large-scale language model to enhance functionality and improve the user interface.

[0002] Computer systems (electronic medical records) for efficiently recording medical instructions (orders) and records have become widely used in medical settings. Furthermore, there has been remarkable progress in artificial intelligence (AI) in recent years, and attention has been focused on large-scale language models (LLMs), which are trained on large volumes of literature and data to enable inquiries and responses in natural language. Prior art documents related to this application include the following:

[0003] Patent Publication No. 2023-523644 Patent Publication No. 2023-73095 Patent Publication No. 5484317 Patent Publication No. 7313757

[0004] When providing medical care, examination plans are created based on the patient's symptoms and examination findings, and treatment plans are created based on the test results, followed by the implementation of treatment and records. Because a wide range of medical professionals, including doctors and nurses, are involved, it goes without saying that electronic medical records are useful for efficient medical activities. However, medical science advances rapidly, with new, highly effective drugs and treatments being developed one after another. Keeping up with the latest medical knowledge places a heavy burden on medical professionals. Medical professionals manually create medical instructions (orders) for tests and procedures, and record their implementation. This requires a huge amount of work, resulting in long working hours. Furthermore, differences in the knowledge and skill levels of medical professionals result in variations in treatment effectiveness.

[0005] The present invention has been made to solve these conventional problems, and its purpose is to provide a system that uses a large-scale language model that has been trained with a large amount of the latest medical knowledge to assist medical professionals in creating order documents and record documents in electronic medical records, thereby supporting efficient medical activities.

[0006] As a means for achieving the above-mentioned objective, the electronic medical record system described in claim 1 is an electronic medical record system that also uses a large-scale language model, and is equipped with an electronic medical record item extraction means that extracts information on items necessary for processing documents of an arbitrarily specified document category from the electronic medical record, and an electronic medical record content to prompt transcription means that provides the extracted electronic medical record content as a prompt for the large-scale language model, and the electronic medical record item extraction means is characterized by having an electronic medical record extracted item management means for each document category that manages a list of electronic medical record items required for each document category.

[0007] The electronic medical record system of claim 2 is characterized in that, in the electronic medical record system of claim 1, it is provided with a means for transcribing the response content to the prompt into the corresponding document in the electronic medical record.

[0008] The electronic medical record system of claim 3 is characterized in that, in the electronic medical record system of claim 1 or 2, it has an electronic medical record additional document creation means for creating at least one of an order document, a record document, or a disease name registration document to be added to the electronic medical record in response to the large-scale language model.

[0009] In the electronic medical record system according to claim 4, in the electronic medical record system according to claim 1 or 2, the large-scale language model has a means for linking with an external system.

[0010] The electronic medical record system of claim 5 is the electronic medical record system of claim 1 or 2, characterized in that the large-scale language model is additionally trained with medical-specific training data, and has a reference data limiting means that refers only to the medical-specific training data when responding.

[0011] The electronic medical record system of claim 6 is characterized in that, in the electronic medical record system of claim 1 or 2, it has a batch processing prompt creation means for inputting multiple prompt contents at once in the prompts of the large-scale language model.

[0012] The electronic medical record system of claim 7 is characterized in that, in the electronic medical record system of claim 1 or 2, the large-scale language model has a document content recommendation means for inputting patient information obtained by the electronic medical record item extraction means for documents of a specified document category as prompt content and requesting recommended content as an answer.

[0013] The electronic medical record system of claim 8 is characterized in that, in the electronic medical record system of claim 1 or 2, the electronic medical record system is provided with an API for each electronic medical record module, and the large-scale language model calls the electronic medical record module via the API to perform processing for the necessary processing.

[0014] The electronic medical record system of claim 9 is characterized in that, in the electronic medical record system of claim 8, it has a series of processing group management means that registers the necessary processing as a series of coherent series of processing groups and makes it possible to specify the series of processing groups for the large-scale language model.

[0015] The electronic medical record system according to claim 1 includes an electronic medical record item extraction means, which extracts information on items required for processing documents of an arbitrarily designated document category from the electronic medical record in an electronic medical record system that also uses a large-scale language model. The system also includes a means for transcribing the electronic medical record contents into a prompt, which provides the extracted electronic medical record contents as a prompt for the large-scale language model.

[0016] The electronic medical record system according to claim 1 includes an electronic medical record extraction item management means for each document category, and therefore the electronic medical record item extraction means manages a list of electronic medical record items required for each document category.

[0017] The electronic medical record system according to claim 2 includes a means for transcribing the contents of the response to the electronic medical record, so that the contents of the response to the prompt are transcribed into the corresponding document in the electronic medical record.

[0018] The electronic medical record system described in claim 3 is equipped with an electronic medical record additional document creation means, and in response to the large-scale language model, creates at least one of an order document, a record document, and a disease name registration document to the electronic medical record as necessary.

[0019] In the electronic medical record system according to claim 4, the large-scale language model has a means for linking with an external system.

[0020] In the electronic medical record system according to claim 5, a reference data limiting means is provided, so that the large-scale language model undergoes reinforcement learning using medical-specific training data, and only the medical-specific training data is referenced when providing an answer.

[0021] The electronic medical record system according to claim 6 includes a batch processing prompt creation means, so that the contents of a plurality of prompts are inputted in a batch in the prompts of the large-scale language model.

[0022] The electronic medical record system of claim 7 is equipped with a document content recommendation means, so that in a large-scale language model, for documents of a specified document category, patient information obtained by the electronic medical record item extraction means is input as prompt content, and recommended content is obtained as an answer.

[0023] In the electronic medical record system according to claim 8, the electronic medical record system is provided with an API for each electronic medical record module, and the large-scale language model performs necessary processing by calling the electronic medical record module via the API.

[0024] In the electronic medical record system according to claim 9, a process series management means is provided, so that necessary processes can be registered as a series of process series, and can be designated as a process series for the large-scale language model.

[0025] 1 is a diagram illustrating a hardware configuration of the present invention. FIG. 1 is an example of a user interface for a large-scale language model. FIG. 2 is a diagram illustrating examples of various document categories in an electronic medical record. FIG. 3 is an example of extending a large-scale language model using an external linkage system. FIG. 4 is an example of data to be additionally trained on a large-scale language model. FIG. 5 is an example of correcting a drug prescription order in an electronic medical record using a large-scale language model. FIG. 6 is an example of verifying a disease name corresponding to a drug prescription order in an electronic medical record. FIG. 7 is an example of verifying a missing disease name and registering an additional disease name in an electronic medical record. FIG. 8 is an explanatory diagram illustrating referring to blood test and radiology test histories from an electronic medical record and ordering a recommended test. FIG. 9 is a diagram illustrating a flow of referring to an electronic medical record using a large-scale language model, placing a necessary additional order, correcting the content based on the results, and recording it in the electronic medical record. FIG. 10 is a diagram of an electronic medical record extracted item management means for each document category. FIG. 11 is a diagram illustrating an example of a batch processing prompt. FIG. 12 is a diagram illustrating an example of prompting input of symptoms, findings, test findings, and disease name, and having the LLM respond with a recommended drug prescription example. FIG. 13 is a conceptual diagram illustrating the LLM taking the initiative in referring to extracted items in an electronic medical record and creating order documents for tests and treatments.

[0026] The electronic medical record system according to the present invention comprises a server device, a database, and a terminal. The server device is a known computer device and comprises an arithmetic unit, a main memory device, an auxiliary memory device, an input device, an output device, and a communication device. The arithmetic unit, the main memory device, the auxiliary memory device, the input device, the output device, and the communication device are connected to each other via an interface. The arithmetic unit comprises a known processor capable of executing an instruction set. The main memory device comprises volatile memory such as RAM capable of temporarily storing the instruction set. The auxiliary memory device comprises non-volatile data storage capable of recording an OS and programs. The data storage may be, for example, an HDD or an SSD. The input device is, for example, a keyboard. The output device is, for example, a display such as an LCD. The communication device comprises a network interface connectable to a network. The server device has means for extracting electronic medical record items, transcribing electronic medical record contents to prompts, managing electronic medical record extracted items by document category, transcribing response contents to electronic medical records, creating additional electronic medical record documents, limiting reference data, creating batch processing prompts, recommending document content, and managing a series of processing groups. The processor of the server device exerts the effects of these means. The database of the present invention may be configured by an auxiliary storage device of the server device, or may be configured by a separate auxiliary storage device independent of the server device. The database stores information handled by the electronic medical record system. The terminal of the present invention, like the server device, has a known computer hardware configuration. The server device, database, and terminal of the present invention can communicate via a network.

[0027] FIG. 1 shows an example of a hardware configuration diagram of the present invention. Because large-scale language models require huge amounts of data and computational resources, they are typically stored on the cloud and connected to a local area network (LAN) within a medical institution via an internet line and a router. An electronic medical record server is located within the medical institution. Doctors, nurses, and other staff members access the electronic medical records and large-scale language models via terminals connected to the LAN. It is also possible to build part or all of the electronic medical record system on the cloud, or, in the case of a lightweight version, to install part or all of the large-scale language model within the medical institution.

[0028] Figure 2 shows an example of a user interface for a large-scale language model (LLM). LLMs are currently being rapidly developed, with numerous models being developed, including ChatGPT (a registered trademark of OpenAI), Bard, LaMDA (a registered trademark of Google), and LLaMA (a registered trademark of Meta). Naturally, user interfaces vary, but typically, as shown in Figure 2, they consist of a box for entering prompts to instruct or inquire of the LLM, a box for displaying responses to those prompts, and a box for displaying a history of prompts and responses as a usage log.

[0029] Figure 3 shows examples of various document categories in an electronic medical record. The electronic medical record contains chronologically recorded records of various observations for each patient, as well as order documents for tests and treatments, created by staff from various professions.

[0030] Figure 4 shows an example of a dedicated LLM for a large-scale language model, using function calling. While the LLM learns large amounts of data to construct a general-purpose language model (base learning model), knowledge in a specific field is not necessarily deep. For this reason, specialized models are constructed by intensively learning additional knowledge in a specific field (see the drug-specific LLM in the figure). Furthermore, the LLM of a general-purpose base learning model is known to be weak in numerical calculations and logical operations. To compensate for this weakness, there are specialized modules such as Wolfram Alpha (see Wolfram Alpha Non-Patent Document 1) (see Numerical Logic Processing System in the figure), both of which respond to prompt queries from the LLM and return answers.

[0031] Furthermore, there is often a delay in acquiring knowledge about recent events. Therefore, for example, a system could query a web search engine for today's weather and retrieve the response. Function calling has been proposed as one approach to solving this problem. This involves using the functions of other search engines, databases, or specialized LLMs via an API (Application Programming Interface). Taking the present invention as an example, while drug efficacy descriptions are stable, drug prices fluctuate annually. While current drug price information can be learned on a board or dedicated LLM, recording and searching in a dedicated database provides faster response and requires less frequent LLM learning updates, reducing costs. Furthermore, using dedicated LLMs in parallel can reduce response times. Furthermore, as described below, by enabling information access from and writing to the electronic medical record system, the LLM can be used to control the electronic medical record system directly.

[0032] Figure 5 shows an example of data to be additionally trained on a large-scale language model in this invention. The interview form for each drug contains dozens of pages of information, including the drug name, dosage, and indications. These descriptions are based on the data used for approval by the Ministry of Health, Labor and Welfare. Disease data includes medical textbooks and disease-specific treatment guidelines compiled by various academic societies. Tests include blood tests, physiological tests such as electrocardiograms, and explanatory books on X-rays, CT scans, and MRI scans. All of these are rigorously reviewed by professionals such as doctors and nurses, ensuring extremely high reliability. In the case of the board LLM, malicious or discriminatory texts have been mixed into some of the large amount of training data, resulting in malicious or discriminatory responses, which has become a problem. Furthermore, "hallucinations," which cause false content to be perceived as fact, reduce the reliability of the responses. These problems can be avoided by limiting the range of training data referenced for answers to highly reliable training data, as shown in this figure (reference data limitation means).

[0033] Figure 6 shows an example of correcting a drug prescription order in an electronic medical record using a large-scale language model. A drug prescription is entered in the drug prescription document in the electronic medical record. As shown in the figure, the prescription content is transcribed into the LLM prompt (a means for transcribing electronic medical record content into a prompt). The LLM provides a response while referring to the content entered in the interview form, etc. If the doctor determines that the correction is appropriate, the response is transcribed into the drug prescription document in the electronic medical record (a means for transcribing response content into the electronic medical record). Note that "transcription into the electronic medical record" not only involves transcribing the response content into the document, but also includes creating a new additional document as needed and transcribing the response content into that newly created document.

[0034] Figure 7 shows an example of verifying disease names corresponding to drug prescription orders in electronic medical records. The drug prescription contents and a list of disease names are extracted from the electronic medical record (electronic medical record item extraction means) and transcribed into a prompt (transcribing electronic medical record contents into a prompt: shown again). The response verifies the drug indication data and the applicable drugs for each disease, and indicates that no additional disease names are required.

[0035] Figure 8 shows an example of verifying missing disease names and registering them in the electronic medical record. The same prompt as in Figure 7 is used. This figure points out that there are no disease names for which Opalmon is suitable. Furthermore, a list of disease names for which Opalmon is suitable is presented, and the doctor is asked to select a disease name. The selected disease name is then transcribed into the electronic medical record (method of transcribing answers into the electronic medical record: shown again), and the disease name is registered. This can reduce the cost of medication and testing (denial of medical fees by insurers) due to missing disease names, which can sometimes exceed several million yen per month.

[0036] Figure 9 is an explanatory diagram for recommending test orders by referencing blood test and radiology history from the electronic medical record. Take the example of a drug prescription for rheumatoid arthritis. Rheumatoid arthritis treatment guidelines recommend blood tests at least once every few months to check for anemia, liver dysfunction, and kidney dysfunction, which are side effects of Rheumatrex, due to bone marrow suppression. Chest X-rays are also recommended approximately once every six months to check for interstitial pneumonia, a side effect of Rheumatrex. The electronic medical record item extraction method extracts the dates of recent blood tests and chest X-rays. If there is a long interval between tests, the system recommends the test and creates an order (electronic medical record additional document creation method). If side effects or a worsening trend are detected, the system recommends discontinuing the drug.

[0037] Figure 10 shows the process of using a large-scale language model, such as those shown in Figures 6 to 9, to reference electronic medical record items, place necessary additional orders, and revise and transcribe the content into the electronic medical record. The target document (in this example, a drug prescription) is selected. The electronic medical record items required for LLM processing of the document are extracted from the electronic medical record and transcribed into the prompt. The electronic medical record items required for LLM processing differ depending on the document category. While prompt transcription can be performed manually each time, as shown in Figure 11, by registering electronic medical record extracted items by document category (electronic medical record extracted item management means for each document category), electronic medical record extracted items can be easily and completely obtained, and the necessary parts can be transcribed into the prompt (electronic medical record content to prompt transcription means). Once a response is received from the LLM, the content of the document is revised (response content to electronic medical record transcription means), or, if necessary, additional documents such as blood test orders or radiology test orders are created (electronic medical record additional document creation means).

[0038] Figure 11 is an explanatory diagram of the electronic medical record extracted item management means for each document category. Items to be transcribed into prompts are managed by the document category in which they are written. As shown in Figures 6 to 9, extracted content can be transcribed into prompts as needed. When ordering an MRI, the items shown in the figure are checked to see if they are written, and if not, the patient or their family members are contacted to confirm. If there is an intracranial aneurysm clip, a pacemaker installed, or metal objects inserted into the body near the target area, the examination is stopped.

[0039] Figure 12 shows an example of a batch processing prompt. In the examples up to now, prompts and answers have been given one step at a time, but if the processing routine is fixed, processing requests for each step can be made in a batch with the prompt. If human intervention is required along the way, a decision can be made or additional documentation can be created, and the processing can continue from there.

[0040] Figure 13 shows an example in which symptoms, findings, test results, and disease name are entered into the prompt, and the LLM responds with a recommended medication prescription. Until now, the contents of the electronic medical record were transcribed into the prompt and queried by the LLM, but the LLM can also generate recommended content as a generative AI. In Figure 13, a recommended medication prescription is requested based on the patient's attributes such as age and gender, symptoms, findings, test result data, and disease name. If the doctor deems it appropriate, it can be transcribed into the electronic medical record. If necessary, recommendations for a chronological document (a so-called clinical path) of multiple prescriptions, such as intravenous drips, rehabilitation, lifestyle guidance, and dietary instructions, can also be requested and transcribed into the electronic medical record.

[0041] Figure 14 is a conceptual diagram in which the LLM takes the lead in referencing extracted items in the electronic medical record and creating test and treatment order documents. In the examples of Figures 6 to 12, a prompt was created based on the electronic medical record document category document, and the LLM was then called. In other words, the electronic medical record category document was the main driver, and the LLM was utilized as needed via the prompt. In contrast, in this diagram, the LLM primarily calls the APIs of each electronic medical record module, acquiring data on items to be referenced, such as symptoms and findings, creating test orders where necessary, estimating and registering disease names based on the obtained test results, and creating order documents for treatment and tests based on the estimated disease name. This system allows for flexible combinations of electronic medical record modules, making it easy to respond to a variety of pathologies. Each electronic medical record module must be accessible from an external system, such as the LLM, via an API. Naturally, each electronic medical record module API requires user eligibility verification.

[0042] Although the above describes an embodiment, the specific configuration of the present invention is not limited to the above embodiment, and design changes that do not deviate from the gist of the invention are also included in the present invention. For example, while the medical field has been used as an example, similar systems can be used in nursing care. While ChatGPT and other systems have been used as examples of LLM, the same applies to other systems, and even systems newly developed in the future will be included in the present invention.

Claims

1. An electronic medical record system that uses a large-scale language model, comprising: an electronic medical record item extraction means for extracting information on items required for processing documents of an arbitrarily specified document category from an electronic medical record; and a transcription means for transcribing the extracted electronic medical record content into a prompt for providing the content as a prompt for a large-scale language model; and the electronic medical record item extraction means having an electronic medical record extracted item management means for each document category for managing a list of electronic medical record items required for each document category.

2. The electronic medical record system according to claim 1, further comprising a transcription means for transcribing answers to said prompts into the corresponding documents in the electronic medical record.

3. An electronic medical record system as described in claim 1 or 2, characterized in that it has an electronic medical record additional document creation means for creating at least one of an additional order document, record document, or disease name registration document to the electronic medical record in response to the large-scale language model.

4. An electronic medical record system according to claim 1 or 2, characterized in that the large-scale language model has a means for linking with an external system.

5. An electronic medical record system as described in claim 1 or 2, characterized in that the large-scale language model is additionally trained with medical-specific learning data, and has a reference data limiting means for referring only to the medical-specific learning data when providing an answer.

6. An electronic medical record system according to claim 1 or 2, further comprising a batch processing prompt creation means for inputting a plurality of prompt contents in a batch in the prompts of said large-scale language model.

7. An electronic medical record system as described in claim 1 or 2, characterized in that in the large-scale language model, a document content recommendation means is provided which inputs patient information obtained by the electronic medical record item extraction means as prompt content for documents of a specified document category, and requests recommended content as an answer.

8. The electronic medical record system according to claim 1 or 2, characterized in that the electronic medical record system is provided with an API for each electronic medical record module, and the large-scale language model calls the electronic medical record module via the API to perform necessary processing.

9. An electronic medical record system according to claim 8, further comprising a process group management means for registering the necessary processes as a series of process groups, and enabling the processes to be specified as a process group for the large-scale language model.

Images