Medical event identification method, device, equipment and storage medium

Abstract

The present disclosure provides a medical event recognition method, device, equipment and storage medium, the method comprising: identifying entities in medical data, obtaining an entity list including multiple types of entities, the multiple types of entities including time entities, subject entities, location entities and event entities; traversing the entity list, obtaining an event list including multiple extracted medical events; wherein for each traversed entity, if the entity is a time entity, the current event time is updated; if the entity is a subject entity, the current event subject is updated; if the entity is a location entity, the current event location is updated; if the entity is an event entity, the event entity is used as the entity content of the current event; and the medical event is obtained based on the current event time, the current event subject, the current event location and the entity content of the current event. The method disclosed in the present disclosure processes medical data through semantic recognition, so that the relationship between medical data is clear and the medical data is fully applied.

Description

Technical Field

[0001] The present disclosure relates to the field of computers, and in particular to a medical event recognition method, device, equipment and storage medium. Background Art

[0002] Medical data is the foundation of medical research, and it supports the research of the entire medical technology. Medical data contains many elements such as person, time, type, action content, etc. For example, the patient underwent a CT examination on March 2, 2021. The person is the patient, the time is March 2, 2021, the type is examination, and the action content is CT examination. Medical data basically records relevant information based on the patient's visit, examination, and treatment. Due to the dispersion of medical data and the very complex relationship, the application of medical data is limited. Summary of the invention

[0003] The present disclosure provides a medical event identification method, device, equipment and storage medium to at least solve the above technical problems existing in the prior art.

[0004] According to a first aspect of the present disclosure, a medical event identification method is provided, comprising:

[0005] Identify entities in the medical data and obtain an entity list, wherein the entity list includes multiple types of entities, and the multiple types of entities include time entities, subject entities, location entities, and event entities;

[0006] Traversing the entity list to extract medical events, and obtaining an event list including the extracted multiple medical events;

[0007] For each entity in the traversed entity list,

[0008] If the entity is a time entity, updating the current event time according to the time entity;

[0009] If the entity is a subject entity, updating the current event subject according to the subject entity;

[0010] If the entity is a location entity, updating the current event location according to the location entity;

[0011] If the entity is an event entity, the event entity is used as the entity content of the current event;

[0012] The medical event is obtained based on the current event time, the current event subject, the current event location and the entity content of the current event.

[0013] In one embodiment, the method further comprises:

[0014] Grouping the events in the event list according to event type to obtain an event type group list, wherein the types of events in each event type group in the event type group list are the same;

[0015] Grouping the events in each event type group in the event type group list by time to obtain an event type time group list;

[0016] If the number of events in the event type time group in the event type time group list is greater than 1, a concurrent relationship is added to all the events in the event type time group.

[0017] In one embodiment, the method further comprises:

[0018] Grouping the events in each event type time group in the event type time grouping list according to the core words to obtain an event type time core word grouping list;

[0019] If the number of events in the event type time core word group in the event type time core word group list is greater than 1, a coreference relationship is added to all the events in the event type time core word group.

[0020] In one embodiment, the method further comprises:

[0021] Traversing the event list;

[0022] If the current event is the first event, an event group is established and the current event is added to the event group;

[0023] If the current event is not the first event, it is matched with the last event group that has been established. If the conditions for joining the last event group are met, the current event is added to the last event group. If the conditions for joining the last event group are not met, a new event group is established, and the current event is added to the new event group to obtain an event grouping list.

[0024] In one possible implementation, when the current event meets the definition of the last event group and is in the same sentence with the events in the last event group, the current event meets the condition of being added to the last event group.

[0025] In one embodiment, the method further comprises:

[0026] The first event group in the event grouping list was selected as a clinical relationship group;

[0027] Traverse the remaining event groups;

[0028] If the traversed event group and the last clinical relationship group do not satisfy the clause condition, the current event group constitutes a new clinical relationship group;

[0029] If the traversed event group does not satisfy the clause sentence condition with the last clinical relationship group, and satisfies the condition for building a clinical relationship with the last clinical relationship group, then the current event group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition to obtain the first clinical relationship group list.

[0030] In one embodiment, the method further comprises:

[0031] Select the first group in the first clinical relationship group list as a clinical relationship group;

[0032] traverse the remaining first clinical relationship group list;

[0033] If the current group traversed and the last clinical relationship group do not satisfy the main clause condition, the current group constitutes a new clinical relationship group;

[0034] If the current group traversed does not meet the main sentence condition with the last clinical relationship group, and meets the condition for building a clinical relationship with the last clinical relationship group, then the current group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition, to obtain the second clinical relationship group list.

[0035] According to a second aspect of the present disclosure, a medical event identification device is provided, the device comprising:

[0036] An identification module, which is used to identify entities in medical data and obtain an entity list, wherein the entity list includes multiple types of entities, and the multiple types of entities include time entities, subject entities, location entities and event entities;

[0037] An extraction module is used to traverse the entity list to extract medical events and obtain an event list including multiple extracted medical events; wherein for each entity in the traversed entity list, if the entity is a time entity, the current event time is updated according to the time entity; if the entity is a subject entity, the current event subject is updated according to the subject entity; if the entity is a location entity, the current event location is updated according to the location entity; if the entity is an event entity, the event entity is used as the entity content of the current event; the medical event is obtained based on the current event time, the current event subject, the current event location and the entity content of the current event.

[0038] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0039] at least one processor; and

[0040] a memory communicatively connected to the at least one processor; wherein,

[0041] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the present disclosure.

[0042] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method described in the present disclosure.

[0043] The medical event recognition method, device, equipment and storage medium disclosed in the present invention can identify entities in medical data through semantic recognition to obtain an entity list; extract medical events from the entity list to obtain an event list including multiple medical events. The event list obtained in the embodiment of the present invention includes all medical events extracted from the medical data, and the relationship between the medical data is clear. Therefore, based on the event list, the medical data can be accurately and quickly applied to medical research and other fields.

[0044] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present disclosure are shown in an exemplary and non-limiting manner, in which:

[0046] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts.

[0047] Figure 1 The schematic diagram shows the implementation process of the medical event identification method of the present disclosure embodiment. Figure 1 ;

[0048] Figure 2 The schematic diagram shows the implementation process of the medical event identification method of the present disclosure embodiment. Figure 2 ;

[0049] Figure 3 A schematic diagram of the implementation code of the medical event identification method according to the embodiment of the present disclosure is shown;

[0050] Figure 4The schematic diagram shows the implementation process of the medical event identification method of the present disclosure embodiment. Figure 3 ;

[0051] Figure 5 A schematic diagram showing the structure of a medical event identification device according to an embodiment of the present disclosure is shown;

[0052] Figure 6 A schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0053] In order to make the purpose, features, and advantages of the present disclosure more obvious and easy to understand, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0054] See also Figure 1 The present disclosure provides a method for identifying a medical event, including:

[0055] Identify entities in the medical data and obtain an entity list, the entity list includes multiple types of entities, the multiple types of entities include time entities, subject entities, location entities and event entities;

[0056] Traversing the entity list to extract medical events, and obtaining an event list including the extracted multiple medical events;

[0057] Among them, for the type of each entity in the traversed entity list,

[0058] If the entity is a time entity, the current event time is updated according to the time entity;

[0059] If the entity is a subject entity, the current event subject is updated according to the subject entity;

[0060] If the entity is a location entity, the current event location is updated according to the location entity;

[0061] If the entity is an event entity, the event entity is the entity content of the current event;

[0062] Medical events are obtained based on the current event time, the current event subject, the current event location and the entity content of the current event.

[0063] The medical event recognition method disclosed in the present invention can identify entities in medical data through semantic recognition to obtain an entity list; extract medical events from the entity list to obtain an event list including multiple medical events. The event list obtained in the embodiment of the present invention includes all medical events extracted from the medical data, and the relationship between the medical data is clear. Therefore, based on the event list, medical data can be accurately and quickly applied to medical research and other fields.

[0064] In the disclosed embodiment, medical data is processed based on medical semantic recognition to extract medical events. The elements related to medical events extracted based on medical semantic recognition mainly include: event type, event entity, event time, event subject, event status, event source, and event location.

[0065] Events are divided into different types. By type, events mainly include drug events, physical examination events, diagnostic events, disease events, test events, radiotherapy events, surgical events, clinical events, operation events, etc. For example, the patient used Bristol-Myers Squibb for treatment on March 5, 2021. This event is classified as a drug event by type. For another example, the patient underwent ovarian cancer cytoreductive surgery on March 5, 2021. This event is classified as a surgical event by type.

[0066] Event entities refer to the specific content of an event. The main event entities include drug entities, inspection entities, diagnostic entities, test entities, gene entities, immunohistochemistry entities, diagnostic entities, clinical findings entities, radiotherapy entities, etc. The content of event entities mainly includes entity type, core words, sentence ID, status, source, and attributes. The attributes include quantitative, qualitative, descriptive words, dimension words, anatomy, directional words, units, etc. The core words of event entities are the most core parts that need to be identified. For example, β2 microglobulin, leukocyturia, hemoglobin concentration, etc. belong to the core words of testing; CK8 / 18, EBERS, etc. belong to the core words of immunohistochemistry; HER2, K-RAS, SLC7A13, etc. belong to the core words of genes; respiratory viruses, filoviridae, poxviridae, etc. belong to the core words of microorganisms. For example, on March 5, 2021, 200 mg of Sintilimab was used for treatment. The event type is a drug event, and the event entity is Sintilimab 200 mg. The event entity contains the core word Sintilimab, the quantity included is 200, and the unit included is mg.

[0067] In real medical data scenarios, especially in current medical history, there may be a doctor who recommends a treatment plan, but the patient or the patient's family disagrees or other reasons do not actually happen. However, this information is more important when statistically analyzing treatment methods. Therefore, all events are set with event status. The event status is the state of the event. The event status includes positive, negative, and uncertain states. Generally, events are positive. Event entities have entity states, which are generally set to the state corresponding to the event entity through the state word in the same clause as the event entity. The event state inherits the state of the event entity. For example, on March 5, 2021, 200 mg of sindilimab was used for treatment, and the status of the event was positive.

[0068] The event source refers to the part of the medical data where the event occurs, such as the current medical history from the inpatient medical record, the diagnosis from the inpatient medical record, the description of the diagnosis and treatment process from the inpatient medical record, etc. The event entity will include the source attribute, and the event source inherits the source of the event entity.

[0069] The event time is the time when the event occurs, mainly referring to the recorded time of the actual event. For example, if 200 mg of Sintilimab was used for treatment on March 5, 2021, the time of the event was March 5, 2021.

[0070] The subject of the event refers to the specific patient in whom the real event occurred. Generally, it refers to the patient himself, but in some cases it is the patient's relative. For example, the patient with the disease described in the family history is a relative of the patient.

[0071] The location of the incident refers to the hospital department where the incident occurred. Generally, it refers to the hospital department of the current case, but in some cases it may also be other hospitals. For example, the current medical history may describe that the diagnosis and treatment were carried out in a certain department of a certain hospital.

[0072] Event recognition refers to the process of medical data eventually forming a medical event, which includes entity recognition of medical data, current event time extraction, current event subject extraction, and current event location extraction. In an exemplary embodiment, before traversing the entity list, current event time extraction, current event subject extraction, and current event location extraction are also performed.

[0073] Medical data includes medical information, such as consultation time, consultation department, consultation age, consultation type, etc.; patient information, such as patient name, gender, date of birth, etc.; specific detailed event source original information of different behaviors, for example: hospitalization records will have current medical history, family history, personal history, physical examination, chief complaint, etc., examination reports will have examination findings, examination conclusions, etc.; discharge records will have occurrence details, diagnosis and treatment process, etc.

[0074] The current event time extraction is to extract the consultation time based on the consultation information, and use the extracted consultation time as the current event time. When the time does not appear in the subsequent entity recognition, the current event time is used as the time when the current event occurs.

[0075] The current event subject extraction is based on the patient information extraction of medical data. It mainly generates event subject information based on the patient's name, gender, and date of birth, and uses the extracted patient name as the current event subject. When the subject does not appear in the subsequent entity recognition, the current event subject is used as the subject of the current event.

[0076] The current event location extraction is also based on the medical information generation, mainly the medical department, to generate the current event location. When the location does not appear in the subsequent entity recognition, the current event location is used as the location where the current event occurred.

[0077] Entity recognition is mainly based on the original information of detailed event sources of different behaviors, such as the current medical history of admission records, examination findings and examination conclusions of examination records, and the output is specific entity information, which includes the core words, types, sentence IDs, status, attributes, entity sources, etc. Entity recognition obtains a list of entity recognition results.

[0078] When extracting events, the event source is the event entity source; the event status is the event entity status; the event type is the event entity type; the event time is the current event time; the event location is the current event location. If the current event location is empty, the default event location is used directly as the current event location. For example, if the event originates from the xx department of xx hospital, the default event location is the xx department of xx hospital. If the event location is identified, the current event location is updated; the event subject is the current event subject. If the current event subject is empty, the default event subject is used as the current event subject. Generally, the patient is the default current event subject. If the event subject is identified, the current event subject is updated. Based on these elements, the complete data of an event in the event list is constructed.

[0079] Event relationship refers to the connection between events. The main event relationships include co-referential relationship, concurrent relationship and clinical relationship.

[0080] A concurrent relationship refers to the occurrence of the same type of events at the same time. During the medical consultation, many similar things were done on the same day or in the same time period. For example, on May 3, 2012, the patient underwent a CT scan and an MRI scan, both of which are examinations and constitute a concurrent relationship.

[0081] In one possible implementation, the medical event identification method of the disclosed embodiment also includes: grouping the events in the event list by event type to obtain an event type grouping list, wherein the types of events in each event type group in the event type grouping list are the same; grouping the events in each event type group in the event type grouping list by time to obtain an event type time grouping list; if the number of events in the event type time group in the event type time grouping list is greater than 1, adding a concurrent relationship to all events in the event type time group.

[0082] A co-referential relationship refers to the relationship between events that have the same event type, event occurrence time, and core words of the event entity. For example, when doctors write medical records, they sometimes copy the previous medical records for convenience, which will result in many medical records with the same information. The events they form should all refer to the same event, so events like this form a co-referential relationship.

[0083] In one possible implementation, the method further includes: grouping the events in each event type time group in the event type time group list according to the core words of the event entity to obtain an event type time core word group list;

[0084] If the number of events in the event type time core word group in the event type time core word group list is greater than 1, a coreference relationship is added to all events in the event type time core word group.

[0085] Clinical relationships are based on the semantics of medical texts and have certain logic. They conform to certain medical clinical laws. In clinical practice, there are event relationships with certain time sequences and certain causal relationships. Based on clinical relationships, we can clearly see the relationship between the treatment, performance, and examination of patients at different stages of diagnosis and treatment. Clinical relationships are a kind of basic data for medical applications, and are widely used in patient diagnosis and treatment path mining, patient treatment method analysis, and intelligent search of medical data. Clinical relationships include many types of relationships. In the exemplary embodiment, clinical relationships can be divided into eight relationships: treatment, discovery after treatment, examination, discovery after examination, diagnosis, discovery after diagnosis, examination after treatment, and diagnosis after treatment.

[0086] The relationship determination of clinical events can be divided into two parts: event grouping and event relationship calculation.

[0087] Event grouping refers to reclassifying the types of events into groups suitable for relationship identification, in order to more conveniently calculate the eight clinical relationships. Events can be divided into manifestation group, treatment group, examination group, and diagnosis group. Manifestations mainly include clinical signs found through physical examinations, laboratory examinations, genetic examinations, etc., symptoms reported by patients to doctors, and diseases diagnosed by doctors. Events in the manifestation group can include clinical discovery events, disease events, immunohistochemical events, genetic events, and test events. Treatment refers to an intervention method for manifestations. The types of events included in the treatment group include drug events, surgical events, and radiotherapy events. Examination refers to physical examinations and laboratory examinations. The examination group includes physical examination events, laboratory examination events, genetic examination events, and immunohistochemical examination events. The diagnosis group contains diagnostic trigger words, mainly diagnostic means events. Event grouping is grouped according to clauses, that is, events and related events before and after are in a clause and meet the definition of the same event grouping. Each medical data includes at least one whole sentence, and each whole sentence includes at least one clause. When identifying specific sentence patterns, it can be identified based on punctuation. The whole sentence is separated by a period, semicolon, exclamation mark and question mark. A whole sentence can be separated into multiple clauses by a comma. For example, in the event list [immunohistochemistry events, test events, drug events, clinical discovery events, disease events], immunohistochemistry events and test events are in a clause and both belong to the performance group, then a performance group is obtained. The performance group contains the elements of the element: the event group type is the performance group, the event group sentence is a clause, and the event group list includes immunohistochemistry events and test events. If the drug event does not meet the same grouping definition as the previous test event and the subsequent clinical discovery event, then the drug event becomes an event group called the treatment group. The event group list of this event group includes the drug event. If the clinical discovery event and the disease event are not in the same clause, although they meet the definition of the performance group, they can only form two groups, and the group name is both performance groups. The event group lists of the two performance groups include clinical discovery events and disease events. The clinical relationship can analyze the relationship between the treatment, manifestation, and examination of patients at different diagnosis and treatment stages. In the embodiment of the present disclosure, using clauses as grouping conditions can improve accuracy.

[0088] In one embodiment, the medical event identification method also includes: traversing the event list; if the current event is the first event, establishing an event group and adding the current event to the event group; if the current event is not the first event, matching it with the last event group established, if the conditions for joining the last event group are met, adding the current event to the last event group, if the conditions for joining the last event group are not met, re-establishing an event group, adding the current event to the re-established event group, and obtaining an event grouping list.

[0089] In one possible implementation, when the current event meets the definition of the last event group and is in the same sentence with the events in the last event group, the current event meets the condition of being added to the last event group.

[0090] The determination of event relationships is the process of matching clinical relationships based on event grouping. The principles of the clinical relationships formed by each event group in the event grouping list according to the temporal relationship and causal relationship are as follows: the manifestation group and the treatment group constitute treatment, the manifestation group and the examination group constitute examination, the manifestation group and the diagnosis group constitute diagnosis, the treatment group and the manifestation group constitute post-treatment findings, the examination group and the manifestation group constitute post-examination findings, the diagnosis group and the manifestation group constitute post-diagnosis findings, the treatment group and the examination group constitute post-treatment examination, and the treatment group and the diagnosis group constitute post-treatment diagnosis.

[0091] In order to ensure the accuracy of event relationships, event relationships have a priority order. In the exemplary embodiment, under the condition of satisfying the temporal relationship, the event group relationship of the clause is calculated first, and then the event group relationship of the whole sentence is calculated; then the event group priority across the whole sentence is calculated. To determine the clinical relationship, the sentence conditions must be met first, and then certain clinical relationship conditions must be met. The clinical relationship conditions are a series of rule configurations formed through medical text analysis and mining. More than 80 rule strategies are defined between event groups, between event groups and event relationship groups, and between event relationship groups, such as Figure 3 is a rule strategy, event_type_list: ["treatment", "treatment"] indicates treatment relationship group and event group treatment, sub_sentences is to form treatment relationship group under a clause condition, and the strategy function used is first_and_current_add. whole_sentences is the whole sentence, and the defined rules are the same as the clauses.

[0092] See also Figure 4, forming a clinical relationship group includes: first selecting the first group in the list as the clinical relationship group, which can be the above-mentioned event group list, or a clinical relationship group list obtained by determining the relationship with the event. For example, when determining the event relationship of the whole sentence, it is based on the clinical relationship group list obtained by determining the event relationship of the sentence; then traversing the remaining group list; when the traversed group and the last clinical relationship group do not meet the sentence condition, the current group constitutes a new clinical relationship group, and the next group is traversed; if the sentence condition is met, and the last clinical relationship group meets the condition for building a clinical relationship, the group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition. For example: the performance group constitutes a clinical relationship group, and the subsequent treatment group and the performance group are in a clause, and meet the clinical relationship of forming treatment, then the name of the last clinical relationship group is treatment, which includes the performance event group and the treatment event group. Continue to traverse an event group. Sentence conditions include clause sentence conditions, whole sentence sentence conditions, and cross-sentence sentence conditions. If the conditions for building a clinical relationship are not met, the current event group forms a new event relationship group, and the next event group is traversed. Finally, the corresponding clinical relationship names are added to the events in the clinical relationship group. If there is a concurrent relationship between the events in the group, the corresponding clinical relationship names are also added to the events of the concurrent relationship; for groups that do not form a clinical relationship, no clinical relationship name is added.

[0093] In one embodiment, the event relationship that satisfies the clause sentence condition is determined, including: selecting the first event group in the event grouping list as a clinical relationship group; traversing the remaining event groups; if the traversed event group and the last clinical relationship group do not satisfy the clause sentence condition, the current event group constitutes a new clinical relationship group; if the traversed event group and the last clinical relationship group do not satisfy the clause sentence condition, and satisfy the condition for building a clinical relationship with the last clinical relationship group, the current event group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition, to obtain the first clinical relationship group list.

[0094] In one embodiment, the event relationship that satisfies the whole sentence condition is determined, including: selecting the first group in the first clinical relationship group list as a clinical relationship group; traversing the remaining first clinical relationship group list; if the traversed current group and the last clinical relationship group do not satisfy the main sentence condition, the current group constitutes a new clinical relationship group; if the traversed current group and the last clinical relationship group do not satisfy the main sentence condition, and satisfy the condition for building a clinical relationship with the last clinical relationship group, the current group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition, to obtain a second clinical relationship group list.

[0095] The method of the disclosed embodiment processes medical data based on semantic recognition and identifies medical events, and can be widely used in mining patient diagnosis and treatment pathways, analyzing patient treatment methods, medical knowledge graphs, auxiliary diagnosis and treatment, and intelligent search of medical data.

[0096] The following is an illustration of the application of the method of the embodiment of the present disclosure in intelligent search of medical data.

[0097] In the medical data intelligent search scenario, the user inputs the sentence they want to search for, and the sentence input by the user is parsed to determine what information the user has specifically inputted, that is, to know the user's intention. During the parsing, the input sentence is subjected to event semantic recognition and relationship mining according to the method of the embodiment of the present disclosure. For example, the user inputs "adverse reactions after chemotherapy for breast cancer", and through semantic recognition, it is found that breast cancer is a disease, post-chemotherapy is a relative time after chemotherapy treatment, and adverse reactions are a clinical type. Based on this, the medical data processed by the method of the embodiment of the present disclosure is searched for breast cancer events, drug events, and clinical adverse reaction events. The relationship between the three is treatment and discovery after treatment. The specific process is as follows:

[0098] Search phrase: Adverse effects of chemotherapy for breast cancer.

[0099] Search intent: Clinical phenomena (adverse reactions) occurring after chemotherapy for breast cancer.

[0100] The medical semantic event parsing of the search phrase is:

[0101] Breast cancer was a disease event;

[0102] After chemotherapy, it was the treatment event and relative time;

[0103] Adverse reactions are clinical events.

[0104] The medical event relationship of the search phrase is parsed as:

[0105] Breast cancer is treated with chemotherapy;

[0106] Chemotherapy and adverse reactions constitute post-treatment findings.

[0107] The medical event temporal relationship of the search phrase is parsed as:

[0108] 1) First, breast cancer (disease event);

[0109] 2) followed by chemotherapy (treatment event);

[0110] 3) Finally, there are adverse reactions (clinical events).

[0111] Example medical data:

[0112] The patient underwent modified radical mastectomy for right breast cancer and invasive breast cancer. The third cycle of AC chemotherapy has been completed. She had grade I gastrointestinal reaction and grade 0 bone marrow suppression. She requested to be discharged today and was allowed to be discharged today.

[0113] The result of the method of the embodiment of the present disclosure performing medical event recognition on the example medical data is:

[0114] “Invasive breast cancer” was the disease event;

[0115] “3rd cycle AC chemotherapy” was a treatment event;

[0116] "Gastrointestinal reaction grade I and bone marrow suppression grade 0" were the adverse reactions of clinical events.

[0117] The medical event recognition results of the sample medical data are:

[0118] Invasive breast cancer can be treated with the third cycle of AC chemotherapy;

[0119] The third cycle of AC chemotherapy was successful with grade I gastrointestinal reaction and grade 0 bone marrow suppression.

[0120] The temporal relationship of medical events identified in the sample medical data is analyzed as follows:

[0121] 1) First, invasive breast cancer (disease event);

[0122] 2) followed by the third cycle of AC chemotherapy (treatment event);

[0123] 3) Finally, there is grade I gastrointestinal reaction and grade 0 bone marrow suppression (clinical events).

[0124] Search medical data matching:

[0125] 1) Search for matches of medical events, see Table 1 below.

[0126] Table 1

[0127] Search intent medical events Medical data matching medical events Breast cancer (disease event) Invasive breast cancer (disease event) Chemotherapy (treatment event) The third cycle of AC chemotherapy (treatment event) Adverse reactions (clinical events) Gastrointestinal reaction grade I, bone marrow suppression grade 0 (clinical event)

[0128] 2) Search for event relationship matching, see Table 2 below.

[0129] Table 2

[0130]

[0131] 3) Event time matching is shown in Table 3 below, where the time sequence is arranged from top to bottom.

[0132] Table 3

[0133] Search intent event chronology Medical data matching event time sequence Breast cancer (disease event) Invasive breast cancer (disease event) Chemotherapy (treatment event) The third cycle of AC chemotherapy (treatment event) Adverse reactions (clinical events) Gastrointestinal reaction grade I, bone marrow suppression grade 0 (clinical event)

[0134] Based on intelligent search of medical data, you can obtain medical maps or perform auxiliary treatments, etc.

[0135] For example, when the user enters breast cancer, the corresponding treatment plan and the adverse reactions of various treatment plans can be found. Among them, breast cancer is a disease event, the treatment plan includes surgical events, radiotherapy events, and drug events, and adverse reactions are clinical events. The relationships constituted include treatment and post-treatment findings, thereby obtaining a medical knowledge graph.

[0136] For auxiliary diagnosis and treatment, in many hospitals or Internet hospitals, some simple diseases can be directly given some suggestions or triage by inputting symptoms. For example: when a user performs auxiliary diagnosis and treatment, he first inputs headache and runny nose, which will be recognized as clinical event semantics. When he inputs that he took Lianhua Qingwen and did not get better, it will be recognized as a drug event. The two constitute a relationship for treatment. By matching the medical data processed by the method of the embodiment of the present disclosure, treatment suggestions or triage are provided.

[0137] The disclosed embodiment provides a medical event identification device, which includes an identification module and an extraction module. The identification module is used to identify entities in medical data and obtain an entity list. The entity list includes multiple types of entities, and the multiple types of entities include time entities, subject entities, location entities and event entities; the extraction module is used to traverse the entity list to extract medical events and obtain an event list including the extracted multiple medical events; wherein for each entity in the traversed entity list, if the entity is a time entity, the current event time is updated according to the time entity; if the entity is a subject entity, the current event subject is updated according to the subject entity; if the entity is a location entity, the current event location is updated according to the location entity; if the entity is an event entity, the event entity is used as the entity content of the current event; and the medical event is obtained based on the current event time, the current event subject, the current event location and the entity content of the current event.

[0138] In one possible implementation, the medical event identification device also includes a determination module, which is used to group the events in the event list by event type to obtain an event type grouping list, wherein the types of events in each event type group in the event type grouping list are the same; group the events in each event type group in the event type grouping list by time to obtain an event type time grouping list; if the number of events in the event type time group in the event type time grouping list is greater than 1, then add a concurrent relationship to all events in the event type time group.

[0139] In one possible implementation, the determination module is also used to group the events in each event type time grouping in the event type time grouping list according to the core words of the event entity to obtain an event type time core word grouping list; if the number of events in the event type time core word grouping in the event type time core word grouping list is greater than 1, a co-reference relationship is added to all events in the event type time core word grouping.

[0140] In one possible implementation, the determination module is also used to traverse the event list; if the current event is the first event, an event group is established and the current event is added to the event group; if the current event is not the first event, it is matched with the last event group that has been established. If the conditions for joining the last event group are met, the current event is added to the last event group; if the conditions for joining the last event group are not met, a new event group is established and the current event is added to the new event group to obtain an event grouping list.

[0141] In one possible implementation, when the current event meets the definition of the last event group and is in the same sentence with the events in the last event group, the current event meets the condition of being added to the last event group.

[0142] In one embodiment, the determination module is also used to select the first event group in the event grouping list as a clinical relationship group; traverse the remaining event groups; if the traversed event group and the last clinical relationship group do not meet the clause sentence condition, the current event group constitutes a new clinical relationship group; if the traversed event group and the last clinical relationship group do not meet the clause sentence condition, and meet the clinical relationship construction condition with the last clinical relationship group, then add the current event group to the last clinical relationship group, and modify the name of the last clinical relationship group to the clinical relationship name defined by the hit strategy condition, and obtain the first clinical relationship group list.

[0143] In one embodiment, the determination module is also used to select the first group in the first clinical relationship group list as a clinical relationship group; traverse the remaining first clinical relationship group list; if the traversed current group and the last clinical relationship group do not meet the main sentence condition, the current group constitutes a new clinical relationship group; if the traversed current group and the last clinical relationship group do not meet the main sentence condition, and meet the clinical relationship construction condition with the last clinical relationship group, then add the current group to the last clinical relationship group, and modify the name of the last clinical relationship group to the clinical relationship name defined by the hit strategy condition, and obtain a second clinical relationship group list.

[0144] The medical event identification device of the embodiment of the present disclosure can implement the methods of the above embodiments, and the description of the above method embodiments can be used to understand and explain the device of the embodiment of the present disclosure. For the purpose of brevity and saving space, it will not be repeated here.

[0145] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device and a readable storage medium.

[0146] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0147] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0148] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0149] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the medical event identification method. For example, in some embodiments, the medical event identification method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or a communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the medical event identification method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the medical event identification method in any other appropriate manner (e.g., by means of firmware).

[0150] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0151] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0152] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0153] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0154] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0155] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0156] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0157] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0158] The above is only a specific embodiment of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present disclosure, which should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A medical event identification method, characterized in that: The method comprises: Identify entities in the medical data and obtain an entity list, wherein the entity list includes multiple types of entities, and the multiple types of entities include time entities, subject entities, location entities, and event entities; Traversing the entity list to extract medical events, and obtaining an event list including the extracted multiple medical events; For each entity in the traversed entity list, If the entity is a time entity, updating the current event time according to the time entity; If the entity is a subject entity, updating the current event subject according to the subject entity; If the entity is a location entity, updating the current event location according to the location entity; If the entity is an event entity, the event entity is used as the entity content of the current event; Obtaining the medical event based on the current event time, the current event subject, the current event location and the entity content of the current event; Traversing the event list; If the current event is the first event, an event group is established and the current event is added to the event group; If the current event is not the first event, it is matched with the last event group that has been established. If the condition for joining the last event group is met, the current event is added to the last event group. If the condition for joining the last event group is not met, a new event group is established, and the current event is added to the newly established event group to obtain an event grouping list. The first event group in the event grouping list was selected as a clinical relationship group; Traverse the remaining event groups; If the traversed event group and the last clinical relationship group do not satisfy the clause condition, the current event group constitutes a new clinical relationship group; If the traversed event group does not satisfy the clause sentence condition with the last clinical relationship group, and satisfies the condition for building a clinical relationship with the last clinical relationship group, then the current event group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition to obtain the first clinical relationship group list.

2. The method according to claim 1, characterized in that The method further comprises: Grouping the events in the event list according to event type to obtain an event type group list, wherein the types of events in each event type group in the event type group list are the same; Grouping the events in each event type group in the event type group list by time to obtain an event type time group list; If the number of events in the event type time group in the event type time group list is greater than 1, a concurrent relationship is added to all the events in the event type time group.

3. The method according to claim 2, characterized in that The method further comprises: Grouping the events in each event type time group in the event type time grouping list according to the core words of the event entity to obtain an event type time core word grouping list; If the number of events in the event type time core word group in the event type time core word group list is greater than 1, a coreference relationship is added to all the events in the event type time core word group.

4. The method according to claim 1, characterized in that: When the current event meets the definition of the last event group and is in the same sentence with the events in the last event group, the current event meets the condition of being added to the last event group.

5. The method according to claim 1, characterized in that The method further comprises: Select the first group in the first clinical relationship group list as a clinical relationship group; traverse the remaining first clinical relationship group list; If the current group traversed and the last clinical relationship group do not satisfy the main clause condition, the current group constitutes a new clinical relationship group; If the current group traversed does not meet the main sentence condition with the last clinical relationship group, and meets the condition for building a clinical relationship with the last clinical relationship group, then the current group is added to the last clinical relationship group, and the name of the last clinical relationship group is modified to the clinical relationship name defined by the hit strategy condition, to obtain the second clinical relationship group list.

6. A medical event identification device, characterized in that: The device comprises: An identification module, which is used to identify entities in medical data and obtain an entity list, wherein the entity list includes multiple types of entities, and the multiple types of entities include time entities, subject entities, location entities and event entities; An extraction module, which is used to traverse the entity list to extract medical events and obtain an event list including multiple extracted medical events; wherein for each entity in the traversed entity list, if the entity is a time entity, the current event time is updated according to the time entity; if the entity is a subject entity, the current event subject is updated according to the subject entity; if the entity is a location entity, the current event location is updated according to the location entity; if the entity is an event entity, the event entity is used as the entity content of the current event; the medical event is obtained based on the current event time, the current event subject, the current event location and the entity content of the current event; A determination module, which is used to traverse the event list; if the current event is the first event, an event group is established and the current event is added to the event group; if the current event is not the first event, it is matched with the last event group that has been established, and if the condition for joining the last event group is met, the current event is added to the last event group; if the condition for joining the last event group is not met, a new event group is established, and the current event is added to the newly established event group to obtain an event grouping list; The determination module is also used to select the first event group in the event grouping list as a clinical relationship group; traverse the remaining event groups; if the traversed event group and the last clinical relationship group do not meet the clause sentence condition, the current event group constitutes a new clinical relationship group; if the traversed event group and the last clinical relationship group do not meet the clause sentence condition, and meet the clinical relationship construction condition with the last clinical relationship group, then add the current event group to the last clinical relationship group, and modify the name of the last clinical relationship group to the clinical relationship name defined by the hit strategy condition, and obtain the first clinical relationship group list.

7. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-5.

Images