WO2021199302A1 - Extraction device and extraction method - Google Patents

Abstract

This extraction device is capable of accessing a medical ontology indicating the relationship between medical terms and a database storing episode map patterns relating to injuries/illnesses abstracted by chronologically connecting medical care details for the injuries/illnesses. The extraction device extracts, in accordance with search conditions based on the medical ontology and a template having an item group including an injury/illness name item, information corresponding to the item group of the template from medical care information indicating medical care provided to an insured person, and outputs specific medical care information regarding the insured person. If an episode map pattern relating to a specific injury/illness corresponding to the reason of an insurance claim by the insured person has been extracted from the database, the extraction device uses the episode map pattern relating to the specific injury/illness and the specific medical care information as a basis to generate an episode map relating to the specific injury/illness in which the medical care details from the episode map pattern relating to the specific injury/illness are specifically stated. The extraction unit extracts information corresponding to the item group of the template from the episode map relating to the specific injury/illness, and outputs medical certificate-equivalent information.

Description

Extractor and extraction method

The present invention relates to an extraction device and an extraction method for extracting data.

In the past, the insured requested a paper medical certificate from a medical institution and submitted it to the insurance company by mail or the like. However, this method requires the insured to visit a medical institution to request and receive a medical certificate. In addition, the medical certificate preparation work is a burden for the medical staff, and it is necessary for the insurance company to contact the medical institution if there are any deficiencies in the items described in the medical certificate. In order to improve this situation, a system that supports the preparation of medical certificates at medical institutions, makes insurance claims using medical fee billing information (receipts), and notifies insurance companies of medical records is already known.

For example, Patent Document 1 discloses a system for automatically identifying and extracting medical conditions and supporting facts from electronic medical records. The system takes formatted text extracted from unstructured electronic medical records and uses the formatted text to represent individual document encounters, where each document has its own document type. The medical condition entity and the rationale fact entity referenced in each of the multiple documents are extracted, and the extracted rationale fact entity in the same document is one or more of the medical ontology or the medical knowledge base. Use to link to individual medical state entities extracted from the same document and provide output data representing the linked rationale facts and medical state entities within the same document.

Patent Document 2 discloses a medical expense payment system that simplifies the payment of medical expenses by the insured and the procedure for claiming insurance payments. In this medical expense payment system, data such as medical expenses and payment methods of the insured person (insured person) is sent from the medical institution terminal to the management center. The management center generates medical information for claiming insurance payment based on the received data and sends it to the insurance company terminal. The insurance company examines the insured by referring to the contract details, and if it corresponds to the payment of medical expenses, sends the payment information including the insurance benefits to the management center. The management center offsets the insurance benefits and medical expenses, calculates the amount to be paid to the medical institution, and requests the financial institution terminal to transfer the money to the medical institution's account. The transfer processing result is sent to the management center, and after the collection registration processing, the receipt file is sent to the medical institution terminal.

Patent Document 3 discloses a system that facilitates the issuance of medical certificates at medical institutions and reduces the burden of procedures for claiming benefits for insured persons. This system has a medical certificate creation correspondence table that accumulates the presence or absence of input of items for each medical certificate type for the items required for medical certificate creation, and a medical certificate that accumulates the medical certificate format data associated with the medical certificate type. Equipped with a format data file, to create a medical certificate, search the medical certificate creation correspondence table based on the medical certificate type in the medical certificate issuance request data, and display only the input items corresponding to the medical certificate form. Display the medical certificate creation screen that was designed to be used, receive the medical certificate creation data, extract the corresponding format data from the medical certificate format data file, and write the medical certificate creation data to display the created medical certificate display screen. Let me. The created electronic medical certificate is encrypted and sent to the insurance company server or relay server in association with the separately encrypted benefit claim.

Non-Patent Document 1 discloses a structured system of clinical medical texts. This structured system targets a discharge summary (a document summarizing the insured's course described at the time of discharge), which is a type of medical record, and extracts descriptions of medication and the side effects caused by it. The modules that make up this system perform language processing such as named entity recognition, notational fluctuation absorption, factuality determination, and relationship extraction.

JP-A-2019-49964 Japanese Unexamined Patent Publication No. 2004-126793 Japanese Unexamined Patent Publication No. 2006-85684

However, in reality, it is not socially implemented except for the medical certificate creation support system that presents the medical certificate input items according to the insurance product that the insured subscribes to. This is because it is generally difficult to disclose all medical records when a medical institution discloses medical information to an insurance company, and manual processing is indispensable to extract only essential items such as medical certificates. The reason is that there is.

An object of the present invention is to improve the convenience of insurance claim examination.

An extraction device that is one aspect of the invention disclosed in the present application is an extraction device having a processor that executes a program and a storage device that stores the program, and has a medical ontology that shows a relationship between medical terms. It is possible to access a database that stores episode map patterns related to the injuries and illnesses that are abstracted by connecting the medical treatment contents of the injuries and illnesses in chronological order. By setting search conditions based on the ontology and extracting information corresponding to the item group of the template from the medical information indicating the medical treatment performed to the insured person according to the search conditions, the insured person is insured. When the first extraction process for outputting the insurer's specific medical information and the episode map pattern for a specific injury or illness corresponding to the insured's insurance claim reason are extracted from the database, the episode map for the specific injury or illness Based on the pattern and the specific medical information, an episode map for the specific injury / illness A generation process for generating an episode map for the specific injury / illness that embodies the medical treatment content of the pattern, and the generation process generated by the generation process. It is characterized by executing a second extraction process of outputting information corresponding to a medical certificate corresponding to a medical certificate by extracting information corresponding to the item group of the template from an episode map relating to a specific injury or illness. ..

According to a typical embodiment of the present invention, it is possible to improve the convenience of insurance claim examination. Issues, configurations and effects other than those described above will be clarified by the description of the following examples.

FIG. 1 is a block diagram showing a configuration example of an extraction system. FIG. 2 is a block diagram showing a hardware configuration example of the extraction device. FIG. 3 is an explanatory diagram showing a configuration example of the medical information DB. FIG. 4 is an explanatory diagram showing a configuration example of a medical ontology. FIG. 5 is an explanatory diagram showing a configuration example of the medical certificate equivalent information template DB. FIG. 6 is an explanatory diagram showing a processing example of the information request unit. FIG. 7 is a flowchart showing an example of a processing procedure of the information request unit. FIG. 8 is an explanatory diagram showing a processing example of the first extraction unit. FIG. 9 is a flowchart showing an example of the processing procedure of the first extraction unit. FIG. 10A is an explanatory diagram showing a processing example 1 of the generation unit. FIG. 10B is an explanatory diagram showing a processing example 2 of the generation unit. FIG. 11 is a flowchart showing an example of a processing procedure of the generation unit. FIG. 12 is an explanatory diagram showing a processing example of the second extraction unit. FIG. 13 is a flowchart showing an example of the processing procedure of the second extraction unit. FIG. 14 is an explanatory diagram showing a processing example of the update unit. FIG. 15 is a flowchart showing an example of a processing procedure of the update unit.

<Configuration example of extraction system>
FIG. 1 is a block diagram showing a configuration example of an extraction system. The extraction system 1 includes an extraction device 100 and an insurance business system 150. The insurance business system 150 is a cooperation destination of the extraction device 100, has an insurance subscriber DB 160, and has a medical certificate equivalent information template DB 170, and executes the business of the insurance company. The insurance policyholder DB 160 is a database that stores information on the insurance policyholder. The medical certificate equivalent information template DB 170 is a database that stores the medical certificate equivalent information template. The medical certificate equivalent information template 171 is a template of the medical certificate equivalent information 141. The extraction device 100 may store the medical certificate equivalent information template DB 170.

The extraction device 100 includes an information request unit 101, a first extraction unit 102, a generation unit 103, a second extraction unit 104, and an update unit 105. When the information request unit 101 receives the medical certificate creation request 111, the information request unit 101 generates a medical information request 112 and an enrollment insurance information request 114. The information request unit 101 searches the medical information DB 110 by the medical information request 112, extracts the medical information 113 from the medical information DB 110, and outputs the medical information 113 to the first extraction unit 102. The information request unit 101 outputs the enrollment insurance information request 114 to the insurance business system 150. The insurance business system 150 searches the insurance member DB 160 by the insurance information request 114, extracts the medical certificate equivalent information template 171 of the corresponding insurance member from the medical certificate equivalent information template DB 170, and outputs it to the first extraction unit 102. do.

The first extraction unit 102 applies the medical information 113 to the medical certificate equivalent information template 171 to extract the specific medical information 121. In addition, the first extraction unit 102 searches the medical ontology 120 using the injury / illness name of the specific medical information 121 as a key, and extracts the related medical ontology 122. The generation unit 103 generates an episode map pattern 131 and an episode map 132 by using the specific medical information 121 and the related medical ontology 122. The second extraction unit 104 extracts the medical certificate equivalent information 141 from the episode map 132. The update unit 105 updates the episode map pattern 131 stored in the episode map pattern DB 130.

<Hardware configuration example of extraction device 100>
FIG. 2 is a block diagram showing a hardware configuration example of the extraction device 100. The extraction device 100 includes a processor 201, a storage device 202, an input device 203, an output device 204, and a communication interface (communication IF) 205. The processor 201, the storage device 202, the input device 203, the output device 204, and the communication IF 205 are connected by the bus 206. The processor 201 controls the extraction device 100. The storage device 202 serves as a work area for the processor 201. Further, the storage device 202 is a non-temporary or temporary recording medium for storing various programs and data. Examples of the storage device 202 include a ROM (Read Only Memory), a RAM (Random Access Memory), an HDD (Hard Disk Drive), and a flash memory. The input device 203 inputs data. The input device 203 includes, for example, a keyboard, a mouse, a touch panel, a numeric keypad, and a scanner. The output device 204 outputs data. The output device 204 includes, for example, a display, a printer, and a speaker. The communication IF205 connects to the network and transmits / receives data.

The information request unit 101, the first extraction unit 102, the generation unit 103, the second extraction unit 104, and the update unit 105 shown in FIG. 1 specifically, for example, processor a program stored in the storage device 202. It is realized by letting 201 execute.

<Database>
FIG. 3 is an explanatory diagram showing a configuration example of the medical information DB 110. The medical information DB 110 is a database generated from an electronic medical record DB (not shown) of a medical institution. The medical information DB 110 stores medical information 113 for each insured person (insurance member) who is an insured person. The medical treatment information 113 is information indicating the contents of the medical treatment (examination and treatment) of the insured person, and has a recording date 301, a medical treatment item 302, and a medical treatment content 303. Each time the insured receives medical treatment, an entry consisting of recording date 301, medical treatment item 302, and medical treatment content 303 is added. The recording date 301 is the date on which the medical treatment item 302 and the medical treatment content 303 are recorded. The medical treatment item 302 indicates the type of medical treatment. The medical treatment content 303 shows the details of the medical treatment received by the insured person. Although not shown, each of the medical information 113 includes information that uniquely identifies the insured person or the insured person.

FIG. 4 is an explanatory diagram showing a configuration example of the medical ontology 120. The medical ontology 120 is a database showing the relationships between medical terms such as disease names, treatments, prescriptions, and surgical procedures. For example, medical terms such as disease names, treatments, and prescriptions are designated as nodes (displayed as rectangles in FIG. 4). It is represented by a graph in which related nodes are connected by a link (in FIG. 4, indicated by a line segment connecting rectangles). The link is set with a degree of relevance indicating the strength of the connection between the nodes at both ends. The larger the relevance value, the stronger the connection between the nodes at both ends.

FIG. 5 is an explanatory diagram showing a configuration example of the medical certificate equivalent information template DB170. The medical certificate equivalent information template DB 170 stores the medical certificate equivalent information templates 171a, 171b, 171c, 171d, ... For each type of medical certificate. For example, the medical certificate equivalent information template 171a for cancer insurance may be, for example, the name of the injury or illness, the date of the first diagnosis of the injury or illness, the examination, the diagnosis, the hospitalization, the length of hospital stay, the date of surgery, the name of the surgery, the prescription, the presence or absence of radiation / hyperthermia, and the application of intractable diseases. The presence or absence is included as an item.

<Processing example of information request unit 101>
FIG. 6 is an explanatory diagram showing a processing example of the information requesting unit 101. FIG. 7 is a flowchart showing an example of a processing procedure of the information requesting unit 101. The information requesting unit 101 acquires the medical certificate creation request 111 from the input device 203, the storage device 202, or the communication IF 205 (step S701). The medical certificate creation request 111 is data for the insured to request the creation of a medical certificate.

The medical certificate creation request 111 includes, for example, the name of the insured person, the name of the insurance company, the insurance policy number, the insurance claim reason 600, the hospital name, the period of the insurance claim reason 600, and the like. The name and insurance policy number are examples of information that uniquely identifies the insured person. The medical certificate creation request 111 may include information that uniquely identifies the insured person other than the name and insurance policy number.

Next, the information requesting unit 101 extracts the medical information request 112 from the medical certificate creation request 111 (step S702). The medical information request 112 is data for requesting medical information 113. The medical information 113 uniquely identifies the insured person by at least the name of the insured person, the insurance company name, the insurance policy number, the insurance claim reason 600, the hospital name, the period of the insurance claim reason 600, etc. Contains information to identify.

Next, the information requesting unit 101 searches the medical information DB 110 using the medical information request 112 (step S703), extracts the medical information 113 corresponding to the medical information request 112 from the medical information DB 110, and extracts the medical information 113. Is output to the first extraction unit 102 together with the insurance claim reason 600 (step S704).

Further, the information requesting unit 101 extracts the enrollment insurance information request 114 from the medical certificate creation request 111 and outputs it to the insurance business system 150 (step S705). The enrollment insurance information request 114 is data for requesting the insurance information enrolled by the insured. The insurance information request 114 includes the name of the insured, the name of the insurance company, the insurance policy number, the reason for claim 600, and the like. The insurance information request 114 may include information that uniquely identifies the insured person other than the name and insurance policy number.

The insurance business system 150 extracts the insurance member information from the insurance member DB 160 with reference to the insurance information request 114 from the information requesting unit 101, and corresponds to the insurance claim reason 600 included in the extracted insurance member information. The medical certificate equivalent information template 171 is extracted from the medical certificate equivalent information template DB 170 and returned to the first extraction unit 102. In the case of this example, since the insurance claim reason 600 is “cancer surgery and hospitalization”, the medical certificate equivalent information template 171a for cancer insurance is extracted.

<Processing example of the first extraction unit 102>
FIG. 8 is an explanatory diagram showing a processing example of the first extraction unit 102. FIG. 9 is a flowchart showing an example of a processing procedure of the first extraction unit 102. The first extraction unit 102 acquires medical information 113 and insurance claim reason 600 from the information request unit 101, and receives a medical certificate equivalent information template 171 from the insurance business system 150 (in this example, a medical certificate equivalent information template for cancer insurance). 171a) is acquired (step S901).

As described above, the medical certificate equivalent information template 171 includes items such as the name of the injury or illness, the date of the first diagnosis of the injury or illness, examination, diagnosis, hospitalization, length of hospital stay, date of surgery, surgery name, prescription, presence or absence of radiation / hyperthermia, and intractable disease. Whether or not it is applied is included. The first extraction unit 102 sets search conditions for each of these items using the medical ontology 120 (step S902).

Specifically, for example, the first extraction unit 102 sets the node of the "disease name" of the medical ontology 120 as a search condition for the "injury and illness name". The first extraction unit 102 sets the "diagnosis" node of the medical ontology 120 as a search condition for "diagnosis". The first extraction unit 102 sets the node of "examination" of the medical ontology 120 as a search condition for "examination".

The first extraction unit 102 sets the node of the "surgery method" of the medical ontology 120 as a search condition for the "surgery name". The first extraction unit 102 sets the “prescription” node of the medical ontology 120 as a search condition for the “prescription”. The first extraction unit 102 sets the node of "treatment" of the medical ontology 120 as a search condition for "presence or absence of radiation / hyperthermia". The first extraction unit 102 sets the link destination node of the node of the “disease name” of the medical ontology 120 as a search condition for “whether or not intractable disease is applied”.

The first extraction unit 102 does not set search conditions for items indicating the time such as "first medical examination date of the injury or illness", "hospitalization period", and "surgery date". In this way, the first extraction unit 102 has a correspondence relationship in advance as to which item of the medical certificate equivalent information template 171 corresponds to which node of the medical ontology 120, and sets search conditions according to this correspondence relationship. , Or not.

Next, the first extraction unit 102 determines whether or not the search has been completed for all the items of the medical certificate equivalent information template 171 (step S903). If it is not completed (step S903: No), the first extraction unit 102 selects one unselected item (step S904) and searches the medical information 113 based on the search condition of the selected item (step S905). .. For example, if the selection items are "injury and illness name", "surgery name", "examination", "diagnosis", "prescription", "presence or absence of radiation / hyperthermia" and "presence or absence of intractable disease application", search conditions are set. Therefore, the first extraction unit 102 searches the medical information 113 using the search conditions. On the other hand, since the search conditions are not set for the "first medical examination date", "hospitalization period", and "surgery date" of the injury or illness, the first extraction unit 102 performs the recording date 301 and the medical treatment item 302 of the medical information 113. Search by referring to.

Then, if there is a search result (step S906: Yes), the process proceeds to step S910, and if there is no search result (step S906: No), the process proceeds to step S907.

For example, when the selected item is "injury / illness name", the search condition is the node of "disease name" in the medical ontology 120. The first extraction unit 102 searches for the medical treatment content 303 of the medical treatment information 113 in the node of the “disease name” and the node of the link destination directly or indirectly connected from the node, and if there is a hit character string, the step S910 is performed. If there is no character string to be hit, the process proceeds to step S907. For example, if there is a "brain tumor" in the "disease name" node, it completely matches the "brain tumor" in the medical treatment content 303, so the process proceeds to step S910. Transition.

Also, if the selection item is "the first medical examination date of the injury or illness", the search condition is not set. In this case, the first extraction unit 102 acquires "2019/7/25", which is the oldest date among the recording dates 301 of the medical information 113, as a search result.

Also, the search condition is not set even when the selected item is "hospitalization period". In this case, the first extraction unit 102 has the medical treatment item 302 "discharged" from "2019/09/10" of the recording date 301 in which the medical treatment item 302 is "hospitalization" in the recording date 301 of the medical information 113. The search results up to "2019/09/20" on the recording date 301 are acquired. If there is no “hospitalization” and “discharge” in the medical treatment item 302, there is no search result (step S906: No), and the process proceeds to step S907.

Also, the search condition is not set even when the selected item is "surgery date". In this case, the first extraction unit 102 acquires "2019/9/13" of the recording date 301 in which the medical treatment item 302 of the medical information 113 is "surgery" as a search result (step S906: Yes), and in step S910. Transition. If there is no “surgery” in the medical item 302, the process proceeds to step S907 with no search result (step S906: No).

Step S906: In the case of No, the first extraction unit 102 searches the medical information 113 for a character string including the same unique expression as the selected item (step S907). Specifically, for example, even if it is determined in step S906 that there is no search result (step S906: No), the first extraction unit 102 considers that there is a search result in the case of notational fluctuation or partial match (step). S908: Yes), the process proceeds to step S910, and if there is no notational fluctuation or partial match, there is no search result (step S908: No), and the process proceeds to step S909.

For example, when the selection item is "prescription", the search condition is the "prescription" node of the medical ontology 120. The first extraction unit 102 searches for the medical treatment content 303 of the medical treatment information 113 at the node of “prescription” and the node of the link destination directly or indirectly connected from the node, and if there is a hit character string, the step S910 is performed. If there is no character string to be hit, the process proceeds to step S907. For example, if there is "analgesic administration" in the linked node of the "prescription" node, it does not completely match the "analgesic administration" of the medical treatment content 303, so the process proceeds to step S907, but the "prescription" node " Since "analgesic administration" is a variation in the notation of "analgesic administration" in the medical treatment content 303, there is a search result (step S908: Yes), and the process proceeds to step S910.

If there is no search result in step S908 (step S908: No), the first extraction unit 102 adds a missing value (Null) to the specific medical information 121 as a value of the selection item (step S909), and returns to step S903. ..

In the case of step S906: Yes or S908: Yes, the first extraction unit 102 adds the search result as the value of the selection item to the specific medical information 121 (step S910), and returns to step S903. When the search for all items is completed in step S903 (step S903: Yes), the first extraction unit 102 adds the insurance claim reason 600 to the specific medical information 121 (step S911), and the specific medical information after the addition. 121 is output to the generation unit 103 (step S911).

Next, the first extraction unit 102 extracts the related medical ontology 122 from the medical ontology 120 and outputs it to the generation unit 103 (step S912) using the injury / illness name of the specific medical information 121 as a key, and ends a series of processes. do. In FIG. 8, from the medical ontology 120, the related medical ontology 122a having “brain tumor” as the node of the disease name and the related medical treatment having “diabetes” as the node of the disease name with the injury / disease names “brain tumor” and “diabetes” as the keys. The ontology 122b is extracted.

In this way, the first extraction unit 102 sets search conditions based on the medical certificate equivalent information template 171 having the item group including the injury / illness name item and the medical ontology 120, and is insured according to the search conditions. The specific medical information 121 of the insured person is output by extracting the information corresponding to the item group of the medical certificate equivalent information template 171 from the medical information 113 indicating the medical treatment performed to the person.

<Processing example of generation unit 103>
FIG. 10A is an explanatory diagram showing a processing example 1 of the generation unit 103. FIG. 10B is an explanatory diagram showing a processing example 2 of the generation unit 103. FIG. 11 is a flowchart showing an example of a processing procedure of the generation unit 103. The generation unit 103 generates the episode map 132 and the episode map pattern 131. The episode map 132 is data embodying the medical treatment contents (nodes such as initial diagnosis, examination (imaging diagnosis), ...) Of the episode map pattern 131. Specifically, for example, episode map 132 is data showing clinical semantic and time-series relationships. As shown in FIGS. 10A and 10B, the episode map 132 includes, as fields, item 1001, content 1002, start date 1003, end date 1004, related injury / illness name 1005, insurance claim target information 1006, and so on. Has. The combination of the values of each field 1001 to 1006 in the same row indicates one episode.

Item 1001 indicates the types of clinically meaningful or time-series episodes. The content 1002 is detailed information of the item 1001 and is acquired from the specific medical information 121. The start date 1003 is the date when the item started, and the end date 1004 is the date when the item ends, which is acquired from the specific medical information 121. The related injury / illness name 1005 is an injury / illness name related to the content 1002, and is obtained from the injury / illness name of the specific medical information 121 or the related medical ontology 122. The insurance claim target information 1006 indicates whether or not the item is subject to an insurance claim.

Episode map pattern 131 is time-series data related to injuries and illnesses that is abstracted by connecting the medical treatment contents of injuries and illnesses (nodes such as initial diagnosis, examination (imaging diagnosis), ...) In chronological order. Specifically, for example, the episode map pattern 131 is time-series data in which episode maps 132 of the same injury or illness are aggregated and patterned, and the patterned node groups subject to insurance claims are connected in the order of appearance. The episode map pattern 131a shown in FIGS. 10A and 10B is a flow of a series of billing nodes such as initial examination ⇒ examination (imaging diagnosis), examination (blood test) ⇒ diagnosis (brain tumor) ⇒ hospitalization ⇒ surgery (craniotomy) ⇒ discharge. Is. The prescription (administration of analgesic) is "Unknown" because it is unknown whether or not it is the subject of the claim.

In FIG. 11, the generation unit 103 acquires the specific medical information 121 and the related medical ontology 122 from the first extraction unit 102 (step S1101).

The generation unit 103 refers to the corresponding information (not shown) in which the insurance claim reason 600 and the injury / illness name are associated with each other, and narrows down the injury / illness name and the related medical ontology 122 according to the insurance claim reason 600 in the specific medical information 121. (Step S1102). In the correspondence information, it is assumed that "cancer surgery and hospitalization", which is the reason for insurance claim 600, corresponds to the injury / illness name "brain tumor". Specifically, for example, in FIGS. 10A and 10B, the insurance claim reason 600 is "cancer surgery and hospitalization". Here, in the specific medical information 121, there is a "craniotomy" as an operation performed. Among the related medical ontology 122, the node having "craniotomy" is the related medical ontology 122a whose injury / disease name is "brain tumor", so that the injury / disease name "brain tumor" and the related medical ontology 122a are narrowed down as search targets. ..

Next, the generation unit 103 searches for the episode map pattern 131 from the episode map pattern DB 130 using the injury / disease name “brain tumor” narrowed down in step S1102 as a key (step S1103). If there is a corresponding episode map pattern 131 (step S1104: Yes), the generation unit 103 extracts the corresponding episode map pattern 131 from the episode map pattern DB 130 (step S1105).

Then, the generation unit 103 generates the episode map 132 based on the extracted episode map pattern 131 and the specific medical information 121 (step S1106). Specifically, for example, the generation unit 103 extracts the nodes included in the extraction episode map pattern 131 in step S1105 as item 1001.

For example, in FIG. 10A, when the episode map pattern 131a is the extracted episode map pattern 131 in step S1105, the nodes "test (imaging diagnosis)", "test (blood test)", and "diagnosis" included in the episode map pattern 131a. (Brain tumor) ”,“ hospitalization ”,“ surgery (craniotomy) ”, and“ prescription (administration of painkiller) ”are extracted as item 1001 of episode map 132.

Since the injury / illness name "brain tumor" is the search target, the "first visit" node is extracted as item 1001: the content of the injury / illness name 1002: brain tumor. The "discharge" node is aggregated as the end date 1004 of the "hospital" node. For the "inspection (imaging diagnosis)" node, the item 1001 is "inspection" and the content 1002 is "imaging diagnosis". The same applies to "test (blood test)", "diagnosis (brain tumor)", "surgery (craniotomy)", and "prescription (analgesic administration)".

Next, the generation unit 103 refers to the specific medical information 121, and if the order of the time series of the nodes extracted as the item 1001 on the time axis of the episode map pattern 131a matches, the insurance claim target information 1006 of the item 1001 Is set to "Yes", and the related injury / illness name 1005 is set to the injury / illness name "brain tumor" of the related medical ontology 122a to be searched. Specifically, for example, when the episode map pattern 131 appears in the time series of node A ⇒ node B ⇒ node C, item A (corresponding to node A) ⇒ item B (corresponding to node B) in the specific medical information 121. ) ⇒ If it is the time series of item C (corresponding to node C), for item B (corresponding to node B), the order of the time series matches between the specific medical information 121 and the episode map pattern 131.

Regarding whether or not the "prescription (painkiller administration)" node is covered by insurance, the episode map pattern 131a is Unknown, but the order in the specific medical information 121 and the order in the episode map pattern 131a are surgery. In order to match the order after (craniotomy) and before discharge, the related injury / illness name 1005 is set to "brain tumor", and the insurance claim target information 1006 is set to "Yes".

Further, the generation unit 103 refers to the specific medical information 121, and the order of the time series of the nodes extracted as the item 1001 on the time axis of the episode map pattern 131a does not match within a predetermined allowable period (for example, 2 weeks). In this case, the insurance claim target information 1006 of the item 1001 is set to "Unknown", and the related injury / illness name 1005 is set to "Unknown".

Specifically, for example, when the episode map pattern 131 appears in the time series of node A ⇒ node B ⇒ node C, item A (corresponding to node A) ⇒ item C (corresponding to node C) in the specific medical information 121. ) ⇒ If it is the time series of item B (corresponding to node B), the order of the time series does not match between the specific medical information 121 and the episode map pattern 131 for item B (corresponding to node B). If the time difference between the item C and the item B is within a predetermined allowable period (for example, 2 weeks), the insurance claim target information 1006 is set to "Unknown".

However, when the name of the injury or illness to be searched narrowed down in step S1002 corresponds to the insurance claim reason 600 in the above-mentioned correspondence information, the generation unit 103 changes the insurance claim target information 1006 from "Unknown" to "Yes". do. For example, in the correspondence information, since "cancer surgery and hospitalization", which is the reason for insurance claim 600, corresponds to the injury / illness name "brain tumor", the related injury / illness name 1005 is changed from "unknown" to "brain tumor", and the insurance The billing target information 1006 is changed from "Unknown" to "Yes".

Further, the generation unit 103 refers to the specific medical information 121, and the order of the time series of the nodes extracted as the item 1001 on the time axis of the episode map pattern 131a does not match outside the predetermined allowable period (for example, 2 weeks). In this case, the insurance claim target information 1006 of the item 1001 is set to "No", and the related injury / illness name 1005 is set to the injury / illness name "diabetes" of the related medical ontology 122b which is not the search target.

Specifically, for example, when the episode map pattern 131 appears in the time series of node A ⇒ node B ⇒ node C, item A (corresponding to node A) ⇒ item C (corresponding to node C) in the specific medical information 121. ) ⇒ If it is the time series of item B (corresponding to node B), the order of the time series does not match between the specific medical information 121 and the episode map pattern 131 for item B (corresponding to node B). If the time difference between item C and item B is outside the predetermined allowable period (for example, 2 weeks), the insurance claim target information 1006 is set to "No".

Further, the generation unit 103 sets the start date 1003 and the end date 1004 with reference to the specific diagnostic information. For example, for item 1001: "name of injury or illness" 1002: "brain tumor", "2019/07/25", which is the first medical examination date of the injury or illness of the specific medical information 121, is set as the start date 1003. The end date 1004 is not set.

Item 1001: Contents of "examination" 1002: Regarding "imaging diagnosis", from "examination: [imaging diagnosis, blood test], (2019/07/25)" in the specific medical information 121, "2019/07/25". Is set to the start date 1003 and the end date 1004. Content 1002: The same applies to "blood test".

Item 1001: Contents of "diagnosis" 1002: Regarding "brain tumor", "2019/08/01" is the start date from "examination: [brain tumor, diabetes], (2019/08/01)" in the specific medical information 121. It is set to 1003 and the end date 1004.

Item 1001: Contents of "surgery" 1002: Regarding "craniotomy", from "surgery date: 2019/09/13" in the specific medical information 121, "2019/09/13" is changed to the start date 1003 and the end date 1004. Set.

Regarding item 1001: "hospitalization", "brain surgery ward" is set to the content 1002 from "hospitalization: brain surgery ward" and "hospitalization period: 2019/09 / 10-2019 / 09/20" in the specific medical information 121. "2019/09/10" is set to the start date 1003, and "2019/09/20" is set to the end date 1004.

Item 1001: Contents of prescription 1002: Regarding administration of analgesics, from "Prescription: Metformin (August 1, 2019-), Analgesics (September 13-20, 2019)" in Specific Medical Information 121, "2019/09" "/ 13" is set to the start date 1003, and "2019/09/20" is set to the end date 1004.

In this way, as shown in FIG. 10A, a new episode map 132a is generated from the specific medical information 121 and the existing episode map pattern 131a, and the process proceeds to step S1108.

Further, in step S1104, when there is no corresponding episode map pattern 131 (step S1104: No), the generation unit 103 newly generates an episode map 132 based on the related medical ontology 122 (step S1107). Specifically, for example, in FIG. 10B, the generation unit 103 indicates each item of the specific medical information 121 (injury / disease name, examination, ... However, the time such as the first medical examination date, hospitalization period, and operation date of the injury / illness. (Excluding items) is set as item 1001 of episode map 132, and related injury / illness name 1005 is assigned.

Specifically, for example, regarding "injury / illness name: [brain tumor, diabetes]" in the specific medical information 121, the generation unit 103 describes the contents of item 1001: "injury / illness name" 1002: "brain tumor" and item 1001: "injury / illness". Content 1002 of "name": Generates "diabetes", assigns "brain tumor" to content 1002: related injury / disease name 1005 of "brain tumor", and assigns "diabetes" to content 1002: related injury / disease name 1005 of "diabetes". Further, each start date 1003 is set to "2019/07/25" from "First diagnosis date of the injury or illness: 2019/07/25" of the specific medical information 121.

Further, regarding "examination: [imaging diagnosis, blood test], (2019/07/25)" of the specific medical information 121, the generation unit 103 describes the content of item 1001: "examination" 1002: "imaging diagnosis" and item. 1001: Content of "test" 1002: Generate "blood test" and refer to related medical ontology 122, content 1002: assign "brain tumor" to related injury and disease name 1005 of "imaging diagnosis", content 1002: "blood" "Brain tumor" is assigned to the related injury / illness name 1005 of "examination". Further, "2019/07/25" is set for each of the start date 1003 and the end date 1004.

Regarding "Diagnosis: [Brain tumor, diabetes], (2019/08/01)" of the specific medical information 121, the generation unit 103 describes the contents of item 1001: "diagnosis" 1002: "brain tumor" and item 1001: ". Content 1002 of "Diagnosis": Generate "Diagnosis" and refer to the related medical ontology 122, Content 1002: Assign "Brain tumor" to the related injury / disease name 1005 of "Brain tumor", Content 1002: Related injury / disease name of "Diagnosis" Assign "diagnosis" to 1005. Further, "2019/08/01" is set for each of the start date 1003 and the end date 1004.

Regarding the "hospitalization: brain surgery ward" of the specific medical information 121, the generation unit 103 generates the content 1002: "brain surgery ward" of the item 1001: "hospitalization", and the "hospitalization period: 2019 /" of the specific medical information 121. With reference to "09 / 10-2019 / 09/20", "2019 / 09/10" is set for the start date 1003, and "2019/09/20" is set for the end date 1004.

Regarding the "surgery name: craniotomy" of the specific medical information 121, the generation unit 103 generates the content 1002: "craniotomy" of the item 1001: "surgery", and the "surgery date: 2019" of the specific medical information 121. With reference to "/ 09/13", "2019/09/13" is set for the start date 1003 and the end date 1004.

Regarding "Prescription: Metformin (2019/8/1 ~), Painkiller (2019/9/13 ~ 20)" of the specific medical information 121, the generation unit 103 describes item 1001: "prescription" content 1002: "Metformin administration" and item 1001: "Prescription" content 1002: Generate "painkiller administration" and refer to the related medical ontology 122, content 1002: "Metformin administration" related injury and disease name 1005 "diabetes" Assignment, content 1002: "Brain tumor" is assigned to the related injury / disease name 1005 of "painkiller administration". In addition, item 1001: "prescription" content 1002: "metformin administration" start date 1003 is set to "2019/08/01", and item 1001: "prescription" content 1002: "painkiller administration". “2019/09/13” is set for the start date 1003, and “2019/09/20” is set for the end date 1004.

Regarding "Presence / absence of radiation / hyperthermia: None" in the specific medical information 121, the generation unit 103 is not generated as an item of the episode map 132 because it does not exist in the node of the related medical ontology 122. The same applies to "whether or not intractable diseases are applied: none".

In this way, as shown in FIG. 10B, when there is no corresponding episode map pattern 131 (step S1104: No), a new episode map 132b is generated from the specific medical information 121, and the process proceeds to step S1108.

After step S1106 or S1107, the generation unit 103 generates the episode map pattern 131 by arranging the generated episode maps 132 in the order of medical treatment (step S1108). When the episode map 132 is generated in step S1106, the episode map pattern 131b is generated for the injury / disease name: brain tumor. Compared with the episode map pattern 131a, in the episode map pattern 131a, the node of the prescription (analgesic administration) is Unknown, which is unknown whether or not it is covered by insurance, but in the episode map pattern 131b, the prescription (analgesic administration). Nodes are subject to insurance claims.

Similarly, when the episode map 132 is generated in step S1107, the episode map pattern 131b is generated for the injury / disease name: brain tumor. Then, the generation unit 103 outputs the generated episode map 132 to the second extraction unit 104, stores the generated episode map pattern 131 in the episode map pattern DB 130 (step S1109), and ends a series of processes.

<Processing example of the second extraction unit 104>
FIG. 12 is an explanatory diagram showing a processing example of the second extraction unit 104. FIG. 13 is a flowchart showing an example of a processing procedure of the second extraction unit 104. The second extraction unit 104 extracts the medical certificate equivalent information 141 from the episode map 132 from the generation unit 103. The medical certificate equivalent information 141 is not the medical certificate itself, but is information necessary for the examination of insurance claims. The medical certificate equivalent information 141 has the same items as the medical certificate equivalent information template 171 (injury / illness name, first diagnosis date of the injury / illness, ...), And the value of the item is set from the episode map 132.

The second extraction unit 104 acquires the episode map 132 from the generation unit 103 (step S1301), and determines whether or not the processing has been completed for all the items 1001 of the episode map 132 (step S1302). When all the items 1001 are not completed (step S1302: No), the second extraction unit 104 selects one unselected item 1001 from the episode map 132 (step S1303). The second extraction unit 104 confirms the insurance claim target information 1006 of the selection item 1001 (step S1304).

If the insurance claim target information 1006 is "Yes" (step S1304: Yes), the process proceeds to step S1306. When the insurance claim target information 1006 is "No" (step S1304: No), the process proceeds to step S1307. When the insurance claim target information 1006 is "Unknown" (step S1304: Unknown), whether or not the degree of relevance between the content 1002 of the selection item 1001 and the insurance claim reason 600 in the related medical ontology 122 is equal to or greater than the threshold value. Judgment (step S1305)

When the degree of relevance is equal to or higher than the threshold value, the second extraction unit 104 adds the content 1002, the start date 1003, and the end date 1004 of the selection item 1001 to the medical certificate equivalent information template 171 (step S1306), and steps S1302. Return to. For example, if the selection item 1001 is "injury / illness name", "brain tumor" of content 1002 is added to the injury / illness name of the medical certificate equivalent information template 171 and "2019/07/25" of the start date 1003 on the first diagnosis date of the injury / illness. "Is added. Further, for example, if the selection item 1001 is "surgery", the "surgery" of the content 1002 is added to the surgery name of the medical certificate equivalent information template 171 and the "surgery" of the start date 1003 (or end date 1004) is added. 2019/09/13 ”is added.

When the process proceeds to step S1307, the second extraction unit 104 discards the selection item 1001 (step S1307) and returns to step S1302. For example, "Metformin administration" is not added to the prescription of the medical certificate equivalent information template 171 because the shape-retaining claim target 1006 of the item 1001 "Prescription" whose content 1002 is "Metformin administration" is "No".

When all the items 1001 are completed in step S1302 (step S1302: Yes), the second extraction unit 104 outputs the medical certificate equivalent information 141 (step S1308), and ends a series of processes.

<Processing example of update unit 105>
FIG. 14 is an explanatory diagram showing a processing example of the update unit 105. FIG. 15 is a flowchart showing an example of a processing procedure of the update unit 105. In FIG. 15, first, the update unit 105 determines whether or not a certain number of medical certificate creation requests 111 have been acquired (step S1501). When a certain number is acquired (step S1501: Yes), the update unit 105 determines whether or not there is an unselected injury / illness name (step S1502).

When there is an unselected injury / illness name (step S1502: Yes), the update unit 105 selects an unselected injury / illness name (step S1503). Then, the update unit 105 acquires all the episode map patterns 131 of the selected injury / illness name from the episode map pattern DB 130 (step S1504). Then, the update unit 105 integrates the acquired episode map pattern 131 to generate an integration result 1400, and totals the frequency of node transitions for each node (step S1505).

The integration result 1400 has an item 1001, a content 1002, a transition source 1403, a transition destination 1404, insurance claim target information 1006, and a frequency 1406. Item 1001 is a node of the episode map pattern 131. The content 1002 is information indicating the node which is the item 1001. The transition source 1403 indicates a transition source node of this node (item 1001). The transition destination 1404 indicates a transition destination node of this node (item 1001). The insurance claim target information 1006 indicates whether or not this node (item 1001) is subject to an insurance claim. The frequency 1406 is a ratio obtained by dividing the number of transitions by which this node (item 1001) has transitioned from the transition source or the transition destination by the number of transitions of all the items 1001.

The update unit 105 generates an episode map pattern 131c of the selected injury / illness name from the items in the entry of the frequency 1406 equal to or higher than the threshold value (step S1506). Then, the update unit 105 registers the generated episode map pattern 131c of the selected injury / illness name in the episode map pattern DB 130 (step S1507), and returns to step S1507. In step S1501, the update unit 105 determines whether or not a certain number of medical certificate creation requests 111 have been newly acquired (step S1501). If it has not been acquired (step S1501: No), the update process ends. As a result, the episode map pattern DB 130 is maintained. In this way, it is possible to improve the accuracy of the episode map 132 in the episode map pattern DB 130.

As described above, according to the extraction device 100 of the present embodiment, the items necessary for the examination by the insurance company can be automatically extracted as the medical certificate equivalent information 141, which improves the convenience of the insurance claim examination. Can be planned.

Further, the extraction device 100 according to the above-mentioned Examples 1 and 2 can also be configured as described in (1) to (10) below.

(1) The extraction device 100 having a processor 201 for executing a program and a storage device 202 for storing the program has a medical ontology 120 showing a relationship between medical terms, and medical treatment contents of injury and illness (initial medical examination, examination (imaging diagnosis)). ), ... Nodes) are connected in chronological order to store the abstracted episode map pattern 131 related to the injury / illness, and the episode map pattern DB 130 and the processor 201 can access the item group including the injury / illness name item. The search conditions are set based on the medical certificate equivalent information template 171 and the medical certificate 120, and the medical certificate equivalent information template 171 is obtained from the medical information 113 indicating the medical treatment performed to the insured person according to the search conditions. Regarding the first extraction process that outputs the insured person's specific medical information 121 by extracting the information corresponding to the item group of, and the specific injury or illness (eg, brain tumor) corresponding to the insured person's insurance claim reason 600. When the episode map pattern 131a is extracted from the episode map pattern DB 130, the medical treatment content of the episode map pattern 131a relating to a specific injury or illness is embodied based on the episode map pattern 131a relating to a specific injury or illness and the specific medical information 121. Diagnosis is made by extracting the information corresponding to the item group of the medical certificate equivalent information template 171 from the generation process for generating the episode map 132a related to the specific injury or illness and the episode map 132a for the specific injury or illness generated by the generation process. The second extraction process of outputting the medical certificate equivalent information 141 corresponding to the document is executed.

(2) In the extraction device 100 of the above (1), the episode map 132a relating to a specific injury or illness (eg, brain tumor) is the medical treatment content (item 1001 and content 1002) performed on the insured person and the specific injury or illness (related). Injury and illness name 1005), insurance claim target information 1006 indicating whether or not the medical treatment content is covered by insurance, and the time when the medical treatment content was implemented (start date 1003, end date 1004) are associated with the data. be.

(3) In the extraction device 100 of the above (2), in the generation process, the processor 201 describes the specific medical treatment content in which the order of the time series matches between the specific medical information 121 and the episode map pattern 131a relating to the specific injury or illness. , The insurance claim target information 1006 is set to the information (Yes) indicating that the insurance claim target information 1006 is to be covered.

(4) In the extraction device 100 of the above (2), in the generation process, the processor 201 matches the order of the time series between the specific medical information 121 and the episode map pattern 131a related to the specific injury or illness within a predetermined allowable period. For a specific medical treatment content that is not specified, the insurance claim target information 1006 is set as information (Time) indicating that it is unknown whether or not the insurance claim target information is applicable.

(5) In the extraction device 100 of the above (2), in the generation process, the processor 201 matches the order of the time series between the specific medical information 121 and the episode map pattern 131a related to the specific injury or illness outside the predetermined allowable period. The insurance claim target information 1006 is set to the information (No) indicating that the insurance claim target information 1006 is not covered by the insurance claim for the specific medical treatment content.

(6) In the extraction device 100 of the above (2), in the first extraction process, the processor 201 performs related medical treatment related to an injury or illness (eg, brain tumor) corresponding to the injury or illness name item obtained from the medical information 113 from the medical ontology 120. In the generation process of extracting the ontology 122, the processor 201 determines the episode map 132b relating to a specific injury or illness based on the related medical ontology 122 when the episode map pattern 131a relating to the specific injury or illness is not extracted from the episode map pattern DB 130. To generate.

(7) In the extraction device 100 of the above (2), in the generation process, the processor 201 sets the medical treatment content (item 1001 and content 1002) and time (start date 1003, end date 1004) of the episode map 132a relating to a specific injury or illness. Based on this, a new episode map pattern 131b relating to a specific injury or illness that is abstracted by connecting the medical treatment contents in time series is generated and registered in the episode map pattern DB 130.

(8) In the extraction device 100 of the above (2), in the second extraction process, the processor 201 uses the medical certificate from the episode map 132a relating to the specific injury or illness based on the insurance claim target information 1006 of the episode map 132a relating to the specific injury or illness. By extracting the information corresponding to the item group of the equivalent information template 171, the medical certificate equivalent information 141 is output.

(9) In the extraction device 100 of the above (8), in the second extraction process, the processor 201 does not cover the insurance claim target information 1006 corresponding to the information corresponding to the item of the medical certificate equivalent information template 171. In the case of the information (No) indicating, the information corresponding to the item of the medical certificate equivalent information template 171 is not extracted from the episode map 132a relating to a specific injury or illness.

(10) In the extraction device 100 of the above (8), in the first extraction process, the processor 201 extracts the related medical ontology 122 related to the injury or illness corresponding to the injury or illness name item obtained from the medical information 113 from the medical ontology 120. In the second extraction process, the processor 201 indicates that it is unknown whether or not the insurance claim target information 1006 corresponding to the information corresponding to the item of the medical certificate equivalent information template 171 is covered by the insurance claim (Unknown). If, from the episode map 132a regarding a specific injury or illness, based on the degree of relevance of the information corresponding to the item of the medical certificate equivalent information template 171 and the insured person's insurance claim reason 600 on the related medical coverage 122. Information corresponding to the item of the medical certificate equivalent information template 171 is extracted.

(11) In the extraction device 100 of the above (1), the processor 201 acquires an episode map pattern group having the same injury or illness from the episode map pattern DB 130, and the medical treatment contents (item 1001, content 1002) included in the episode map pattern group. ) Appearance pattern (transition source 1403, transition destination 1404), frequency 1406 is calculated, and a new episode map pattern 131c for injury or illness is generated based on the appearance pattern (transition source 1403, transition destination 1404) and frequency 1406. Then, the update process registered in the episode map pattern DB 130 is executed.

The present invention is not limited to the above-described embodiment, but includes various modifications and equivalent configurations within the scope of the attached claims. For example, the above-described examples have been described in detail in order to explain the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to those having all the described configurations. Further, a part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Further, the configuration of another embodiment may be added to the configuration of one embodiment. In addition, other configurations may be added, deleted, or replaced with respect to a part of the configurations of each embodiment.

Further, each of the above-described configurations, functions, processing units, processing means, etc. may be realized by hardware by designing a part or all of them by, for example, an integrated circuit, and the processor 201 performs each function. It may be realized by software by interpreting and executing the program to be realized.

Information such as programs, tables, and files that realize each function is recorded in a memory, hard disk, storage device such as SSD (Solid State Drive), or IC (Integrated Circuit) card, SD card, DVD (Digital Versailles Disc). It can be stored on a medium.

Also, the control lines and information lines show what is considered necessary for explanation, and do not necessarily show all the control lines and information lines necessary for implementation. In practice, it can be considered that almost all configurations are interconnected.

Claims (12)

1. An extraction device having a processor that executes a program and a storage device that stores the program.
   It is possible to access a medical ontology that shows the relationship between medical terms and a database that stores episode map patterns related to the injuries and illnesses that are abstracted by connecting the medical treatment contents of the injuries and illnesses in chronological order.
   The processor
   Search conditions are set based on a template having an item group including an injury / illness name item and the medical ontology, and from the medical information indicating medical treatment performed to the insured person according to the search conditions, the template The first extraction process that outputs the specific medical information of the insured person by extracting the information corresponding to the item group, and
   When an episode map pattern relating to a specific injury or illness corresponding to the insured person's insurance claim reason is extracted from the database, the identification is based on the episode map pattern relating to the specific injury or illness and the specific medical information. Generation processing to generate an episode map related to the specific injury or illness that embodies the medical treatment content of the episode map pattern related to the injury or illness of
   The second extraction process of outputting the medical certificate equivalent information corresponding to the medical certificate by extracting the information corresponding to the item group of the template from the episode map relating to the specific injury or illness generated by the generation process.
   An extraction device characterized by performing.
2. The extraction device according to claim 1.
   The episode map relating to the specific injury or illness includes the medical treatment content performed on the insured person, the specific injury or illness, insurance claim target information indicating whether or not the medical treatment content is covered by insurance, and the medical treatment. Data that associates the time when the content was implemented,
   An extraction device characterized by that.
3. The extraction device according to claim 2.
   In the generation process, the processor sets the insurance claim target information as the insurance claim target for a specific medical content in which the order of the time series matches between the specific medical information and the episode map pattern related to the specific injury or illness. Set the information to indicate that
   An extraction device characterized by that.
4. The extraction device according to claim 2.
   In the generation process, the processor describes the insurance claim target information for a specific medical treatment content in which the order of the time series does not match between the specific medical treatment information and the episode map pattern related to the specific injury or illness within a predetermined allowable period. Is set in the information indicating that it is unknown whether or not the insurance is covered.
   An extraction device characterized by that.
5. The extraction device according to claim 2.
   In the generation process, the processor describes the insurance claim target information for a specific medical treatment content in which the time series order does not match between the specific medical treatment information and the episode map pattern related to the specific injury or illness outside a predetermined allowable period. Is set as the information indicating that the insurance is not covered by the insurance claim.
   An extraction device characterized by that.
6. The extraction device according to claim 2.
   In the first extraction process, the processor extracts from the medical ontology the related medical ontology related to the injury or illness corresponding to the injury or illness name item obtained from the medical information.
   In the generation process, the processor generates an episode map for the particular injury or illness based on the related medical ontology when the episode map pattern for the particular injury or illness is not extracted from the database.
   An extraction device characterized by that.
7. The extraction device according to claim 2.
   In the generation process, the processor connects the medical treatment contents in chronological order and abstracts the medical treatment contents based on the medical treatment contents and the timing of the episode map related to the specific injury or illness. And register it in the database,
   An extraction device characterized by that.
8. The extraction device according to claim 2.
   In the second extraction process, the processor extracts information corresponding to the item group of the template from the episode map related to the specific injury / illness based on the insurance claim target information of the episode map relating to the specific injury / illness. Outputs the information equivalent to the medical certificate.
   An extraction device characterized by that.
9. The extraction device according to claim 8.
   In the second extraction process, when the processor is information indicating that the insurance claim target information corresponding to the information corresponding to the item of the template is not the insurance claim target, the episode map relating to the specific injury or illness is used. Do not extract the information corresponding to the item of the template,
   An extraction device characterized by that.
10. The extraction device according to claim 8.
    In the first extraction process, the processor extracts from the medical ontology the related medical ontology related to the injury or illness corresponding to the injury or illness name item obtained from the medical information.
    In the second extraction process, when the processor is information indicating that it is unknown whether or not the insurance claim target information corresponding to the information corresponding to the item of the template is the insurance claim target, the information is described. Based on the degree of relevance of the information corresponding to the item of the template and the insurance claim reason of the insured person on the related medical ontology, the item corresponding to the item of the template is applied from the episode map relating to the specific injury or illness. Extract information,
    An extraction device characterized by that.
11. The extraction device according to claim 1.
    The processor
    An episode map pattern group having the same injury or illness is acquired from the database, a frequency is calculated for each appearance pattern of the medical treatment content included in the episode map pattern group, and the injury or illness is based on the appearance pattern and the frequency. Update process, which generates a new episode map pattern and registers it in the database.
    An extraction device characterized by performing.
12. An extraction method executed by an extraction device having a processor that executes a program and a storage device that stores the program.
    The extraction device has access to a medical ontology showing the relationship between medical terms and a database that stores episode map patterns related to the injuries and illnesses, which are abstracted by connecting the medical treatment contents of the injuries and illnesses in chronological order.
    The extraction method is
    The processor
    Search conditions are set based on a template having an item group including an injury / illness name item and the medical ontology, and from the medical information indicating medical treatment performed to the insured person according to the search conditions, the template The first extraction process that outputs the specific medical information of the insured person by extracting the information corresponding to the item group, and
    When an episode map pattern relating to a specific injury or illness corresponding to the insured person's insurance claim reason is extracted from the database, the identification is based on the episode map pattern relating to the specific injury or illness and the specific medical information. Generation processing to generate an episode map related to the specific injury or illness that embodies the medical treatment content of the episode map pattern related to the injury or illness of
    The second extraction process of outputting the medical certificate equivalent information corresponding to the medical certificate by extracting the information corresponding to the item group of the template from the episode map relating to the specific injury or illness generated by the generation process.
    An extraction method characterized by performing.

Images