Method, device, computer equipment and storage medium for generating medical review

Abstract

The embodiments of this specification provide a method, apparatus, computer equipment and storage medium for generating a medical review. It includes: extracting key information including primary key information and secondary key information from the medical report text for the target drug; performing correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain the correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions; generating the correlation evaluation results between the target drug and the adverse drug reaction event based on the correlation analysis results and the preset correlation evaluation rules; generating a medical review report based on the correlation evaluation results. By extracting key information from the medical report text and generating a medical review report in combination with the preset medical review knowledge base and correlation evaluation rules, the quality and standardization of the medical review report are improved.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of pharmacovigilance, and specifically to methods, devices, computer equipment and storage media for generating medical reviews. Background Art

[0002] The medical review of individual safety reports of clinical studies plays an important role in the risk management of drugs and the safety of drug use by the public. Existing medical review reports are mainly written manually by medical reviewers based on individual safety reports of clinical studies, supplemented by reference to relevant medical literature. However, due to the differences in professional levels among different personnel, the medical reviews written are uneven, the quality cannot be guaranteed, and a large amount of medical literature needs to be consulted during the writing process. For the common knowledge between the medical reviews of individual safety reports of different clinical studies, it may be necessary to search the same literature multiple times, resulting in low efficiency. Summary of the invention

[0003] In view of this, multiple embodiments of this specification are dedicated to providing a method, apparatus, computer device and storage medium for generating a medical review, so as to improve the standardization of medical review reports and the work efficiency of medical reviewers.

[0004] The implementation mode of this specification proposes a method for generating a medical review, which is applied to a first terminal, and includes: receiving a medical report text for a target drug sent by a second terminal; extracting key information in the medical report text; wherein the key information includes primary key information describing an adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions; based on the evaluation dimensions, performing correlation analysis on the primary key information and the secondary key information to obtain correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions; wherein the correlation analysis results include the existence of correlation or the absence of correlation; based on the correlation analysis results and a preset correlation evaluation rule, generating a correlation evaluation result between the target drug and the adverse drug reaction event; and generating a medical review report according to the correlation evaluation result.

[0005] The implementation mode of this specification proposes a medical review generation device, which is applied to a first terminal and includes: a medical report text receiving module, which is used to receive a medical report text for a target drug sent by a second terminal; a key information acquisition module, which is used to extract key information in the medical report text; wherein the key information includes primary key information describing an adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions; a correlation analysis module, which is used to perform correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain a correlation analysis result between the primary key information and the secondary key information on different evaluation dimensions; wherein the correlation analysis result includes the existence of correlation or the absence of correlation; a correlation evaluation module, which is used to generate a correlation evaluation result between the target drug and the adverse drug reaction event based on the correlation analysis result and a preset correlation evaluation rule; and a medical review report generation module, which is used to generate a medical review report according to the correlation evaluation result.

[0006] An embodiment of the present specification provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the embodiment is implemented.

[0007] The embodiments of this specification provide a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described in the embodiments is implemented.

[0008] The implementation method of this specification realizes the automatic filling of medical review reports by extracting key information required for filling templates and correlation analysis from individual safety reports of clinical studies, and generating the final medical review report in combination with a preset medical review knowledge base and correlation evaluation rules, thereby improving the standardization of medical review reports. The relevant literature of adverse drug reaction events is determined by keyword search, and the causes of adverse drug reactions corresponding to the adverse drug reaction events in the relevant literature are added to the preset medical review knowledge base, thereby improving the quality and interpretability of the medical review report, and avoiding the search for duplicate literature data, thereby improving the work efficiency of medical reviewers. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 Shown is a schematic diagram of interactions between different ends in a scenario example provided by an implementation method.

[0010] Figure 2 Shown is a schematic diagram of interactions between different ends in a scenario example provided by an implementation method.

[0011] Figure 3 Shown is a schematic flow chart of a method for generating a medical review report provided in one embodiment.

[0012] Figure 4 Shown is a schematic diagram of correlation evaluation rules provided by an implementation method.

[0013] Figure 5 Shown is a schematic diagram of a medical review report generating device provided by one embodiment.

[0014] Figure 6 Shown is a schematic diagram of a computer device provided by an embodiment. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the solution of this specification, the technical solution in the implementation of this specification will be clearly and completely described below in conjunction with the drawings in the implementation of this specification. Obviously, the described implementation is only a part of the implementation of this specification, not all of the implementations. Based on the implementation in this specification, all other implementations obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0016] See also Figure 1 and Figure 2 This specification provides a scenario example of a medical review report generation system, which may include a client and a server. The user may be a medical reviewer in the medical field who needs to generate a medical review report based on the information in the clinical research case safety report.

[0017] The medical reviewer first accepts the clinical research case safety report of the medical review report sent from the second terminal to the first terminal and inputs the medical review report template to the first terminal. Among them, the medical review template includes multiple slots, each slot can contain attribute tags that need to be filled in. Then, the client sends the clinical research case safety report to the server and requests the return of the medical review report. After receiving the clinical research case safety report, the server first extracts key information from the clinical research case safety report based on the attribute information included in the medical review template slot, including the main key information describing the adverse drug reaction event and the secondary key information describing the cause of the adverse drug reaction event.

[0018] After the key information is determined, the server can use the description of adverse drug reactions in the key information as the main key information and the cause of adverse drug reactions as the secondary key information. Then, the server will determine whether the main key information and the secondary key information are related or not in the medical report and the medical review knowledge base. Among them, the medical review knowledge base includes the target drug, the adverse reactions of the target drug, and the causes of the adverse reactions of the target drug. When using the medical review knowledge base to determine the correlation between the main key information and the secondary key information, the server can input the secondary key information and the cause of the adverse reaction into the similarity calculation model to calculate the similarity between the secondary key information and the cause of the adverse reaction. Then, the server will set the secondary key information corresponding to the cause of the adverse reaction whose maximum value in the similarity is greater than or equal to the preset threshold and the result of the correlation analysis of the main key information as related. In addition, the server can set the result of the correlation analysis between the main key information and the secondary key information whose maximum value in the similarity is less than the preset threshold as unrelated.

[0019] After obtaining the correlation analysis results of the main key information and the secondary key information of different dimensions, the server can use the correlation analysis results and the corresponding correlation degree data in the preset correlation evaluation as the correlation evaluation results between the target drug and the adverse drug reaction. Subsequently, the server can fill the correlation analysis results and the correlation evaluation results of the key information, the main key information and the secondary key information into the corresponding event overview template, the correlation evaluation template and the evaluation conclusion template according to the attribute tags corresponding to the key information to obtain the adverse drug reaction event overview report, the correlation evaluation report and the evaluation conclusion report, thereby generating a medical review report including the adverse drug reaction event overview report, the correlation evaluation report and the evaluation conclusion report. Finally, the server will return the medical review report to the client. After receiving the medical review report, the client presents the medical review report to the medical review personnel through the display. Of course, when the medical review report extraction system is installed in the client and the corresponding medical review knowledge base is downloaded, the medical review report can be directly generated in the client and presented to the medical review personnel through the display. After proofreading the automatically generated medical review report, the medical review personnel send it to the sponsor of the target drug as the material for applying for drug marketing.

[0020] The above is only an example scenario provided in this specification and is not intended to limit the present invention. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0021] The implementation method of this specification provides a medical review report generation system. The medical review report generation system can report the client and the server. The client can be an electronic device with network access capability. Specifically, for example, the client can be a desktop computer, a tablet computer, a laptop computer, a smart phone, a digital assistant, a smart wearable device, a shopping guide terminal, a television, a smart speaker, a microphone, etc. Among them, smart wearable devices include but are not limited to smart bracelets, smart watches, smart glasses, smart helmets, smart necklaces, etc. Alternatively, the client can also be software that can be run in the electronic device. The server can be an electronic device with certain computing and processing capabilities. It can have a network communication module, a processor, and a memory, etc. Of course, the server can also refer to software running in the electronic device. The server can also be a distributed server, which can be a system with multiple processors, memories, network communication modules, etc. operating in collaboration. Alternatively, the server can also be a server cluster formed by several servers. Alternatively, with the development of science and technology, the server can also be a new technical means that can realize the corresponding functions of the implementation method of the specification. For example, it can be a new form of "server" based on quantum computing.

[0022] See also Figure 3 The present specification provides a method for generating a medical review report, which is applied to a first terminal and includes the following steps.

[0023] Step S110: receiving a medical report text for the target drug sent by the second terminal.

[0024] In some embodiments, in order to ensure the safety of the medical report text, the medical report text for the target drug is sent to the medical reviewer by the medical worker who conducts the clinical research of the target drug in the research institution. Therefore, the medical reviewer needs to obtain the medical report text sent by the second terminal before writing the medical review report.

[0025] The first terminal is a terminal device used by medical reviewers. The first terminal may be a desktop computer, a tablet computer, a laptop computer, a smart phone, a digital assistant, a smart wearable device, etc., and is used to receive the medical report text for the target drug sent by the second terminal, process the medical report text, and generate a medical review report. The medical reviewer may be a project manager, a pharmacovigilance specialist, a pharmacovigilance medical specialist / manager / director, etc., who is responsible for the target drug.

[0026] The second terminal is a terminal device used by medical workers in clinical trials, and can be used to send safety reports of subjects participating in the target drug trial, including the diagnosis process, treatment process, and treatment results, to the first terminal. The second terminal can be a desktop computer, tablet computer, laptop computer, smart phone, digital assistant, smart wearable device, etc.

[0027] Step S120: extracting key information from the medical report text; wherein the key information includes primary key information describing the adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions.

[0028] The medical report text includes medical data and non-medical data, and the types are various and the sentences are lengthy. If all the data in the medical report text is directly matched with the content that needs to be filled in the slot of the medical review template, the amount of calculation is large and it takes a lot of time. Therefore, the key information in the medical report text can be determined first, and then the key information can be matched with the content that needs to be filled in the slot of the medical review template.

[0029] The medical report text is a research case safety report in the field of clinical trials. Specifically, for example, in a drug clinical research event, the research case safety report may include the patient's diagnosis process, the patient's treatment process and the patient's treatment results. Of course, the research case safety report may also include relevant information such as the patient's medical history, complications, etc.

[0030] The key information includes the information that needs to be filled in the medical review template. In a key information extraction process, the information that needs to be filled in the medical review template includes the patient's diagnosis information, the patient's treatment process information, the description of the patient's adverse reaction events, etc., then it is necessary to extract the corresponding information in the medical report text as key information, and the key information describing the adverse drug reaction events in the key information is used as the main key information, and the key information describing the cause of the adverse drug reaction events is used as the secondary key information. Specifically, for example, the key information extracted from the medical report text includes: breast malignant tumor, malignant pleural effusion, lung excitation malignant tumor, pyroximate, nausea, vomiting, diarrhea, then nausea, vomiting, diarrhea will be used as the main key information, and breast malignant tumor, malignant pleural effusion, lung excitation malignant tumor will be used as the secondary key information.

[0031] The main key information is the description of the adverse drug reaction event. Specifically, for example, the medical report text includes "the patient was admitted to the hospital with a diagnosis of breast malignancy accompanied by malignant pleural effusion, and vomiting and diarrhea occurred after treatment with pyroxithromycin", then the main key information in this sentence is "vomiting" and "diarrhea".

[0032] The secondary key information is the cause of the adverse drug reaction. Specifically, for example, the medical report text includes "the patient was admitted to the hospital with a diagnosis of breast malignancy accompanied by malignant pleural effusion, and vomiting and diarrhea symptoms occurred after treatment with pyroninib", then the secondary key information in this sentence is "malignant pleural effusion" and "pyroninib".

[0033] The evaluation dimension is used to determine whether there is a correlation between a drug and an adverse drug reaction event. Specifically, for example, in the dimension of correlation between combined medication and adverse reactions, the main key information is "diarrhea", and the secondary key information is "aspirin" and "penicillin". After analysis, it is found that "diarrhea" is correlated with aspirin, but not with "penicillin". Therefore, in the dimension of correlation between combined medication and adverse reactions, the correlation between combined medication and adverse reactions is that there is a correlation.

[0034] The method for extracting key information from the medical report text can be to first perform word segmentation processing on the medical report text to obtain the word segmentation result of the medical report text and the corresponding part of speech. The word segmentation result of the medical report text is extracted as key information based on the part of speech and the content that needs to be filled in the medical review template, and each key information includes the attribute label corresponding to the content that needs to be filled in the medical report text. Of course, the corresponding content can also be extracted from the medical report text in the order of the content that needs to be filled in the medical review template and filled into the slot of the medical review template.

[0035] Step S130: performing correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions; wherein the correlation analysis results include the existence of correlation or the absence of correlation.

[0036] In order to obtain the correlation analysis results between the target drug and adverse drug reactions, it is necessary to analyze the correlation between adverse drug reaction events and the causes of adverse drug reaction events. Therefore, the correlation analysis between the main key information and the secondary key information can be performed.

[0037] The obtained correlation analysis results between the main key information and the secondary key information in different evaluation dimensions can be combined with the context of the main key information and the secondary key information in the medical review report to determine the correlation analysis results of the main key information and the secondary key information in different evaluation dimensions. Of course, the correlation analysis results between the main key information and the secondary key information, in which the main key information and the secondary key information have a corresponding relationship in the preset medical review knowledge base, can also be set as the existence of correlation. Specifically, for example, the text of the medical report includes "The subject has been taking the target drug Biroxithamide 480mg bid regularly since the day of enrollment on 2021-03-16. During the period of taking the target drug, nausea, vomiting, and diarrhea occurred many times, and recovered after symptomatic treatment." The correlation analysis result between the last medication time and the adverse drug reaction is that there is a correlation. For another example, in the text of the medical report, "After the hospitalization period, omeprazole sodium for injection and compound amino acid (18AA) injection were given for symptomatic supportive treatment." In the medical review knowledge base, when searching for adverse reaction event reports such as "nausea", "vomiting", and "diarrhea", "omeprazole sodium" is the cause of the adverse reaction event "vomiting", and "compound amino acid (18AA) injection" is not the cause of these three adverse reaction events. Therefore, the result of the association analysis between "omeprazole sodium" and adverse reaction events is related, and the result of the association analysis between "compound amino acid (18AA) injection" and adverse reaction events is not related.

[0038] Step S140: generating a correlation evaluation result between the target drug and the adverse drug reaction event based on the correlation analysis result and a preset correlation evaluation rule.

[0039] In the medical review report, it is necessary to generate a correlation evaluation between the target drug and the adverse drug reaction event. Therefore, when the preset correlation evaluation rules are the same as the association analysis results in different dimensions, the corresponding correlation degree data in the correlation evaluation can be used as the correlation evaluation result between the target drug and the adverse drug reaction event.

[0040] See also Figure 4 . The correlation evaluation rule is to determine the degree of correlation between the target drug and the adverse drug reaction event. Specifically, for example, in the correlation evaluation rule, when the last medication time is related to the adverse drug reaction event, the adverse reaction event is a known adverse reaction event of the target drug, the adverse reaction event does not occur after stopping taking the study medication, and the adverse reaction event occurs again after taking the drug again, the correlation evaluation result between the target drug and the adverse drug reaction event is positively correlated.

[0041] The correlation evaluation result is used to evaluate the degree of correlation between the target drug and the adverse drug reaction event. Specifically, for example, in the correlation analysis results between the primary key information and the secondary key information in different dimensions, if the last medication time is related to the adverse drug reaction event, the adverse reaction event is a known adverse reaction event of the target drug, the adverse reaction event does not occur after stopping taking the study drug, and the adverse reaction event occurs again after taking the drug again, then according to the above correlation evaluation rules, the correlation evaluation result between the target drug and the adverse drug reaction event is positively correlated.

[0042] Step S150: Generate a medical review report based on the correlation evaluation result.

[0043] The medical review report needs to determine the correlation between the target drug and the adverse drug reaction event. Therefore, the key information and correlation evaluation template can be filled into the slot of the evaluation conclusion template in the corresponding medical review template based on the corresponding attributes in the slot of the medical review template, thereby obtaining a medical review report including the evaluation conclusion report.

[0044] In some embodiments, the step of extracting key information from the medical report text may include: obtaining a preset medical review template; wherein the medical review template includes multiple slots; wherein the slots include attributes of information that need to be filled in; and extracting information corresponding to the attributes in the medical report text as key information.

[0045] Correctly extracting key information from medical report texts can, on the one hand, shorten the computer's running time and thus improve efficiency, and on the other hand, improve the accuracy of filling key information into the corresponding slots of the medical review template. Therefore, named entity recognition and / or short text classification methods can be used to extract key information from medical report texts based on the attributes corresponding to the slots of the medical review template.

[0046] The named entity recognition is to search for information of the attribute in the medical report text according to the attribute corresponding to the slot of the medical review template. Specifically, for example, the medical review template includes "This case [age] years old, [sex] sex, [ethnicity] subject." Then, the information corresponding to the attributes can be searched in the medical report text according to the three attributes of "age", "sex" and "ethnicity", which are "54", "female" and "Han nationality" respectively.

[0047] In some embodiments, the step of extracting key information from the medical report text may include: performing word segmentation processing on the medical report text to obtain the word segmentation result of the medical report text and the part of speech of the word segmentation result; and extracting the information required to be filled in the medical review template from the word segmentation result as key information based on the part of speech.

[0048] In some cases, the medical report text can be segmented first to obtain the segmentation results and parts of speech, and then the content required to be filled in the medical review template can be extracted from the segmentation results according to the parts of speech and the content, which can improve the accuracy of key information extraction to a certain extent.

[0049] The method of word segmentation processing can be to perform maximum matching of the medical report text in a medical dictionary to obtain a medical word segmentation result and a corresponding part of speech. Subsequently, the content in the medical report text that is not matched in the medical dictionary can be input into a general word segmentation tool to obtain a word segmentation result of non-medical words and a corresponding part of speech. For example, the content in the medical report text that is not matched in the medical dictionary can be input into a Jieba word segmentation tool to obtain a word segmentation result of non-medical words and corresponding part of speech information. Finally, the medical word segmentation result and the corresponding part of speech and the non-medical word segmentation result and the corresponding part of speech can be used as the result of the medical report text after word segmentation processing.

[0050] The part of speech of the word segmentation result is the part of speech pre-defined in the medical dictionary or the general word segmentation tool. Specifically, for example, the part of speech corresponding to "vomiting" in the medical dictionary is "symptom", then "symptom" is used as the part of speech of the word "vomiting". For another example, the part of speech corresponding to "breast malignancy" in the medical dictionary is "disease name", then "disease name" is used as the "part of speech of breast malignancy". For another example, the result predicted by the Jieba word segmentation tool for the non-medical word "2019-10-02" is "date", then "date" is used as the part of speech of "2019-10-02".

[0051] The medical review template includes slot information for filling in key information, and each slot has attribute tags that need to fill in key information. Specifically, for example, the medical review template contains the sentence "The subject's adverse event [SAE name] has a reasonable time correlation from the last medication date [last medication date] [presence / non-existence]". Among them, "[ ]" is the slot of the medical review template, and "SAE name", "last medication date", and "presence / non-existence" are the attribute information corresponding to the key information that needs to be filled in each slot.

[0052] In some embodiments, the step of extracting key information from the medical report text may further include: adding attribute tags to the key information; wherein the attribute tags correspond to the attributes of the information that needs to be filled in the slots of the medical review template; accordingly, the step of generating a medical review report based on the correlation evaluation results may include: filling the key information and the correlation evaluation results into the correlation evaluation template according to the attribute tags to obtain an evaluation conclusion report; and generating a medical review report based on the evaluation conclusion report.

[0053] In some cases, different key information in the medical report text corresponds to the same part of speech, and when filling the key information into the corresponding slot of the medical review template, it is also necessary to fill in the corresponding key information according to the attributes of each slot. Therefore, by adding corresponding attribute tags to the key information, the key information with the same attribute tags as those defined in the medical review template slot can be filled into the medical review template slot of the same drink.

[0054] The method of adding attribute labels to the key information can be filled in according to the context in the medical report text. Specifically, for example, the medical report text includes the sentence "The subject experienced nausea level 2 and vomiting level 2 again on 2021-05-29, and has been taking 5 mg of Tropisetron Hydrochloride Tablets orally every day since 2021-05-29", then the key information "Oral Tropisetron Hydrochloride Tablets" corresponds to the attribute label "Combined Medication", and the attribute labels corresponding to the two key information "Nausea" and "Vomiting" are "Adverse Drug Reaction".

[0055] In some embodiments, after the step of extracting key information from the medical report text for the target drug, the step may also include: obtaining clinical research process information of the medical report text; wherein the clinical research process information includes at least one of the following: patient diagnosis process information, patient treatment process information, and patient treatment effect analysis process information; filling the key information corresponding to the clinical research process information into an event overview template to obtain a drug adverse reaction event overview report; accordingly, the step of generating a medical review report based on the correlation evaluation results may include: generating a medical review report based on the drug adverse reaction event overview report and the evaluation conclusion report.

[0056] In some cases, the entire development process information of the drug clinical trial needs to be filled in the medical review report template as a summary of adverse drug reaction events. Therefore, the clinical research process information of the medical report text can be obtained to fill in the corresponding event summary template to obtain a summary of adverse drug reaction events.

[0057] The clinical research process information of the medical report text is obtained by obtaining information related to the patient's diagnosis process information, the patient's treatment process information, and the patient's treatment effect in the medical report text. Specifically, for example, the medical report text includes "Subject JIYA, female, 52 years old, Han nationality, left breast cancer surgery for more than 2 years, multiple metastases for more than 1 year, admission diagnosis: targeted treatment of malignant tumors, breast malignant tumors, secondary malignant tumors of lymph nodes". Then, the relevant content of "admission diagnosis" in the medical report text can be used as the patient's diagnosis process.

[0058] In some embodiments, the method for generating a medical review report may further include: filling the correlation analysis results between the primary key information and the secondary key information and between the primary key information and the secondary key information into a correlation analysis template to obtain a correlation analysis report; correspondingly, the step of generating a medical review report based on the correlation evaluation results may include: generating a medical review report based on the adverse reaction event overview report, the correlation analysis report and the evaluation conclusion report.

[0059] In some cases, it is also necessary to conduct a correlation analysis between the adverse drug reactions and the causes of the adverse drug reactions in the medical review report. Therefore, the correlation between the target drug and the adverse drug reaction can be determined based on the results of the correlation evaluation and the evaluation conclusion template in the medical review template.

[0060] The association analysis report includes the results of association analysis in multiple dimensions. Specifically, for example, the association analysis template includes three dimensions of association analysis report templates: association analysis between the last time of study medication and adverse reaction events, association analysis between adverse reaction events and known adverse reaction events of study medication, and association analysis between adverse reaction events after patients stop using study medication. After determining that there is a correlation between the association analysis between adverse reaction events and known adverse reaction events of study medication and the association analysis between adverse reaction events after patients stop using study medication, the association analysis report can be: there is a reasonable time correlation between the last time of study medication and adverse reaction events; the adverse reaction event is a known adverse reaction event; no adverse reaction events occur after the patient stops using study medication.

[0061] The method for generating an evaluation result report based on the correlation evaluation result and the evaluation conclusion template can be determined according to the correlation analysis results of different evaluation dimensions. Specifically, for example, if an adverse drug reaction event is related to the last medication time, then the target drug is related to the adverse drug reaction event. For another example, when the adverse drug reaction event is related to the last medication time, the adverse drug reaction event is not a known adverse drug reaction event of the study drug, the adverse reaction event still exists after the study drug is stopped, and the adverse reaction event disappears after the study drug is used again, the adverse drug reaction event is related to the combined medication, then the target drug is not related to the adverse drug reaction event.

[0062] In some embodiments, the evaluation dimension includes at least one of the following: an analysis of the association between the last time of study medication and adverse events; or, an analysis of the association between adverse events and known adverse events of study medication; or, an analysis of the association between patients stopping the use of study medication and adverse events; or, an analysis of the association between patients stopping the use of study medication and re-using study medication and adverse events; or, an analysis of the association between patients using other drugs during the use of study medication and adverse events; or, an analysis of the association between drugs with the same pharmacological toxicity as the target drug and adverse events; or, an analysis of the association between patients' previous / combined medical history and adverse events; or, an analysis of the association between complications of the patient's disease and adverse events.

[0063] When performing correlation analysis between adverse drug reaction events and the causes of adverse drug reaction events, the analysis can be performed based on different evaluation dimensions. Therefore, it is possible to determine whether the target drug is the cause of the adverse drug reaction event through correlations on a variety of different evaluation dimensions. Specifically, for example, by analyzing the correlation between the patient's last medication time and the adverse reaction event dimension, the correlation evaluation between the target drug and the adverse drug reaction can be obtained.

[0064] In some embodiments, the dimensions of the association analysis include multiple dimensions, characterized in that the step of performing association analysis on the main key information and the secondary key information based on the evaluation dimensions to obtain the correlation analysis results between the main key information and the secondary key information in different evaluation dimensions may include: obtaining the correlation analysis results between the main key information and the secondary key information in different evaluation dimensions; using the correlation analysis results and the corresponding correlation degree data in the preset correlation evaluation rules as the correlation evaluation results between the target drug and the adverse drug reaction event; wherein the correlation evaluation rules include the correlation analysis results in different evaluation dimensions and their corresponding correlation degree data.

[0065] The results of the correlation analysis between the target drug and the adverse drug reaction events need to be filled in the medical review report template. Since different target drugs correspond to different correlation evaluation rules, before evaluating the correlation between the target drug and the adverse drug reaction events, the correlation evaluation rules need to be pre-configured. Then, the correlation evaluation results between the target drug and the adverse drug reaction can be generated based on the correlation evaluation rules and the correlation analysis results of the primary key information and the secondary key information.

[0066] The correlation analysis of the main key and secondary key information on the different evaluation dimensions may include: correlation analysis between the last time of study medication and adverse reaction events, correlation analysis between adverse reaction events and known adverse reaction events of study medication, correlation analysis between the time after the patient stops using the study medication and adverse reaction events, correlation analysis between the patient's use of the study medication again after stopping using the study medication and adverse reaction events, correlation analysis between the patient's use of other drugs during the use of the study medication and adverse reaction events, correlation analysis between drugs with the same pharmacological toxicity as the target drug and adverse reaction events, correlation analysis between the patient's previous / combined medical history and adverse reaction events, and correlation analysis between complications of the patient's disease and adverse reaction events. Specifically, for example, if the time between the patient's last medication and the occurrence of the adverse reaction event is 35 days, which is greater than the set date range of 30 days, then there is no reasonable time correlation between the last time of study medication and the adverse reaction event.

[0067] The method for generating the correlation evaluation result between the target drug and the adverse drug reaction event can be based on the correlation analysis results of the main key information and the secondary key information in different dimensions. When the correlations in different dimensions defined in the correlation evaluation rules are the same, the corresponding correlation degree data in the correlation evaluation rules can be used as the correlation evaluation result between the target drug and the adverse drug reaction event. Of course, the correlation evaluation result between the main key information and the secondary key information in different dimensions can also be determined by the cumulative sum of the correlations in the correlation analysis results of the main key information and the secondary key information in different dimensions and the preset threshold range.

[0068] In some embodiments, the step of performing correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain the correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions may include: determining the correlation analysis results between the primary key information and the secondary key information according to the context of the primary key information and the secondary key information in the medical report text.

[0069] In some cases, the results of the correlation analysis between the main key information and the secondary key information can be obtained by analyzing the text of the medical report. Therefore, the main key information and the secondary key information can be obtained by analyzing the context of the medical report text. Specifically, for example, the text of the medical report includes: "The subject began to take the target drug 480 mg bid of pyroximate regularly on 2021-03-16, the day of enrollment, and has been taking the target drug for many times. Nausea, vomiting, and diarrhea occurred during the period of taking the target drug, and recovered after symptomatic treatment. The subject experienced nausea 2 and vomiting 2 again on 2021-05-29. From 2021-05-29 to date, he took 5 mg of tropisetron hydrochloride tablets orally every day, and from 2021-06-01 to date, he took 20 mg of omeprazole enteric-coated capsules orally every day, and the symptoms did not relieve." This shows that the results of the correlation analysis between the time of adverse drug reactions and re-stimulation are related.

[0070] In some embodiments, the step of performing association analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain the association analysis results between the primary key information and the secondary key information on different evaluation dimensions may also include: determining the adverse drug reaction event description corresponding to the primary key information in a preset medical review knowledge base as the target adverse drug reaction event description; wherein the medical review knowledge base includes target drugs, adverse drug reaction events of target drugs, and causes of target adverse drug reaction events; calculating the similarity between the adverse drug reaction event cause corresponding to the target adverse drug reaction event description in the medical review knowledge base and the secondary key information; when the maximum value in the similarity is greater than or equal to a preset threshold, setting the association analysis result between the secondary key information and the primary key information as having an association relationship. When the maximum value in the similarity is less than a preset threshold, setting the association analysis result between the secondary key information and the primary key information as not having an association relationship.

[0071] In some embodiments, the correlation analysis result between the primary key information and the secondary key information cannot be obtained by analyzing the medical report text. Therefore, it is necessary to perform the correlation analysis with the help of the medical review knowledge base. Of course, since the primary key information and the secondary key information in the medical report text may have slightly different descriptions, but the meanings expressed are the same. Therefore, the correlation between the primary key information and the secondary key information can be set as related if the similarity between the secondary key information and the cause of the adverse reaction in the medical review knowledge base is greater than a preset threshold.

[0072] The medical review knowledge base includes target drugs, descriptions of adverse drug reactions, and causes of adverse drug reactions. Specifically, for example, the target drug "Biloxifen" has a corresponding adverse reaction description; the adverse reaction description also has a corresponding adverse reaction cause. Specifically, for example, the medical report text includes "The subject has been taking the target drug Biloxifen 480 mg bid regularly since the day of enrollment on 2021-03-16. During the period of taking the target drug, nausea, vomiting, and diarrhea occurred many times. After the hospitalization, omeprazole sodium injection and compound amino acid (18AA) injection were given symptomatic supportive treatment." Then, the method for determining the correlation between the main key information and the secondary key information in the medical review knowledge graph can be to take "nausea", "vomiting" and "diarrhea" as the main key information, and "omeprazole sodium" and "compound amino acid (18AA) injection" as the secondary key information. By calculating the maximum similarity between "omeprazole sodium" and the causes of adverse drug reactions in the medical review knowledge base as 100%, which is greater than the preset threshold of 80%, the correlation between "omeprazole sodium" as a combined medication and adverse reactions is considered relevant.

[0073] In some cases, the correlation between the primary key information and the secondary key information is irrelevant. Therefore, the results of the correlation analysis between the secondary key information and the primary key information whose maximum similarity of adverse reactions in the medical review knowledge base is less than the preset threshold can be set as irrelevant. Specifically, for example, the text of the medical report includes "The subject has been taking the target drug pyroximate 480 mg bid regularly since 2021-03-16, the day of enrollment. During the period of taking the target drug, nausea, vomiting, and diarrhea occurred many times. After the hospitalization, omeprazole sodium injection and compound amino acid (18AA) injection were given symptomatic supportive treatment." Then, the method for determining the correlation between the main key information and the secondary key information in the medical review knowledge graph can be to take "nausea", "vomiting" and "diarrhea" as the main key information, and "omeprazole sodium" and "compound amino acid (18AA) injection" as the secondary key information. By calculating the maximum similarity between "compound amino acid (18AA) injection" and the causes of adverse drug reactions in the medical review knowledge base as 50%, which is less than the preset threshold of 80%, the correlation between "compound amino acid (18AA) injection" as a combined medication and adverse reactions is irrelevant.

[0074] In some embodiments, the correlation analysis of the primary key information and the secondary key information to obtain the correlation analysis result between the primary key information and the secondary key information may also include: constructing a medical review knowledge base including target drugs, adverse drug reactions of target drugs, and causes of adverse drug reactions of target drugs. Of course, a medical review knowledge graph may also be constructed based on target drugs, adverse drug reactions of target drugs, and causes of adverse drug reactions of target drugs.

[0075] By constructing a medical review knowledge base, we can make up for the deficiency of only stating facts in the medical report text, but not analyzing the correlation between adverse drug reaction events and the causes of adverse drug reaction events. In addition, due to the limited knowledge level of medical workers when writing medical report texts, they cannot fully determine the correlation between adverse drug reaction events and the causes of adverse drug reaction events described. By constructing a medical review knowledge base, the target drug, the description of the adverse reaction event of the target drug, and the cause of the adverse reaction event of the target drug are included in the medical review knowledge base, which is convenient to improve the accuracy of the correlation evaluation between the main key information and the secondary key information in the medical review report, and can further affect the evaluation results of the correlation between the target drug and the adverse reaction of the target drug.

[0076] In some embodiments, the correlation analysis of the primary key information and the secondary key information to obtain the correlation analysis result between the primary key information and the secondary key information may also include: searching for relevant literature on adverse drug reaction events in a medical literature library; determining the adverse drug reaction event and the cause corresponding to the adverse drug reaction event in the relevant literature; and storing the cause corresponding to the adverse drug reaction event in a medical review knowledge base corresponding to the adverse drug reaction event as the cause of the adverse drug reaction event.

[0077] As drug clinical trials proceed, the number of adverse drug reaction event reports may increase, and many unknown adverse drug reaction events may become known adverse drug reaction events. Therefore, the adverse reactions of target drugs can be updated to the medical review knowledge base through literature search, thereby further improving the accuracy of the correlation evaluation between the main key information and the secondary key information in the medical review report.

[0078] See also Figure 5 In some embodiments, a medical review report generating device may be provided, which may include: a medical report text receiving module, a key information acquisition module, a correlation analysis module, a correlation evaluation module, and a medical review report generating module.

[0079] A key information acquisition module, used to extract key information from the medical report text; wherein the key information includes primary key information describing the adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions;

[0080] A correlation analysis module, configured to perform correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain correlation analysis results between the primary key information and the secondary key information in different evaluation dimensions; wherein the correlation analysis results include the existence of correlation or the absence of correlation;

[0081] A correlation evaluation module, used to generate a correlation evaluation result between the target drug and the adverse drug reaction event based on the correlation analysis result and a preset correlation evaluation rule;

[0082] A medical review report generating module is used to generate a medical review report according to the correlation evaluation result.

[0083] The specific functions and effects achieved by the medical review report generation device can be explained in reference to other embodiments of this specification, and will not be repeated here. The various modules in the medical review report generation device can be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0084] See also Figure 6 In some implementations, a computer device may be provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the method steps in the implementation are implemented.

[0085] In some embodiments, a computer-readable storage medium may be provided, on which a computer program is stored. When the computer program is executed by a processor, the method steps in the embodiments are implemented.

[0086] Those of ordinary skill in the art can understand that all or part of the processes in the implementation method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the implementation methods of the methods described. Among them, any reference to memory, storage, database or other media used in the various implementations provided in this specification may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0087] It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0088] The descriptions of the various embodiments of this specification are made in a progressive manner. Different embodiments focus on describing the parts that are different from other embodiments. After reading this specification, a person skilled in the art can know that the various embodiments in this specification, as well as the various technical features disclosed in the embodiments, can be combined in more ways. To make the description concise, not all possible combinations of the various technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0089] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0090] The multiple embodiments in this specification emphasize the parts that are different from other embodiments, and the various embodiments can be explained in comparison with each other. Any combination of the multiple embodiments in this specification by technicians in the relevant field based on general technical knowledge is included in the disclosure scope of this specification.

[0091] The above description is only an implementation method of this case and is not intended to limit the scope of protection of the claims of this case. For those skilled in the art, this case may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this case should be included in the scope of the claims of this case.

Claims

1. A method for generating a medical review, applied to a first terminal, characterized in that: include: receiving a medical report text for a target drug sent by a second terminal; Extracting key information from the medical report text; wherein the key information includes primary key information describing the adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions; Based on the evaluation dimensions, correlation analysis is performed on the primary key information and the secondary key information to obtain correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions; wherein the correlation analysis results include the existence of correlation or the absence of correlation; Generating a correlation evaluation result between the target drug and the adverse drug reaction event based on the correlation analysis result and a preset correlation evaluation rule; generating a medical review report according to the relevance evaluation results; The step of extracting key information from the medical report text includes: Obtaining a preset medical review template; wherein the medical review template includes a plurality of slots; wherein the slots include attributes of information to be filled in; Extracting information corresponding to the attribute in the medical report text as key information; The step of extracting key information from the medical report text further includes: adding attribute tags to the key information; wherein the attribute tags correspond to the attributes of the information to be filled in the slots of the medical review template; Accordingly, the step of generating a medical review report according to the correlation evaluation result includes: Fill the key information and the relevance evaluation result into the relevance evaluation template according to the attribute label to obtain an evaluation conclusion report; Generate a medical review report based on the evaluation conclusion report; Wherein, after the step of extracting key information from the medical report text, the method further includes: Acquire clinical research process information of the medical report text; wherein the clinical research process information includes at least one of the following: patient diagnosis process information, patient treatment process information, and patient treatment effect analysis process information; Filling the key information corresponding to the clinical research process information into the event summary template to obtain a drug adverse reaction event summary report; Accordingly, the step of generating a medical review report according to the correlation evaluation result includes: Generate a medical review report based on the adverse drug reaction event summary report and the evaluation conclusion report; Wherein, the method further comprises: Fill the main key information and the secondary key information as well as the correlation analysis results between the main key information and the secondary key information into the correlation analysis template to obtain a correlation analysis report; Accordingly, the step of generating a medical review report according to the correlation evaluation result includes: generating a medical review report according to the adverse drug reaction event summary report, the correlation analysis report and the evaluation conclusion report; The step of performing correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions includes: Determine the results of the correlation analysis between the primary key information and the secondary key information according to the context of the primary key information and the secondary key information in the text of the medical report; The step of performing correlation analysis on the primary key information and the secondary key information to obtain a correlation analysis result between the primary key information and the secondary key information further includes: Determining the adverse drug reaction event description corresponding to the main key information in a preset medical review knowledge base as the adverse drug reaction event description of the target drug; wherein the medical review knowledge base includes the target drug, the adverse reaction event of the target drug, and the cause of the adverse reaction event of the target drug; Calculating the similarity between the cause of the adverse drug reaction event corresponding to the target adverse drug reaction event description in the medical review knowledge base and the secondary key information; When the maximum value of the similarities is greater than or equal to a preset threshold, the result of the association analysis between the secondary key information and the primary key information is set as an association relationship; When the maximum value of the similarities is less than a preset threshold, the result of the association analysis between the secondary key information and the primary key information is set to indicate that there is no association relationship.

2. The method according to claim 1, characterized in that The evaluation dimensions include at least one of the following: Correlation analysis between the last time of study medication and adverse events; or, Association analysis between adverse events and known adverse events of the study medication; or Analysis of the association between patients stopping the study medication and adverse events; or, Analysis of the association between the use of study drugs and adverse events after patients stop using the study drugs; or Analysis of the association between the use of other drugs and adverse events during the period of study medication; or Association analysis between drugs with the same pharmacological toxicity as the target drug and adverse events; or, Analysis of the association between the patient's past / combined medical history and adverse events; or, Analysis of the association between complications of patients' diseases and adverse events.

3. The method according to claim 2, characterized in that The step of performing correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain correlation analysis results between the primary key information and the secondary key information on different evaluation dimensions includes: Obtain the correlation analysis results of the main key information and the secondary key information in different evaluation dimensions; The correlation analysis result and the corresponding correlation degree data in the preset correlation evaluation rule are used as the correlation evaluation result between the target drug and the adverse drug reaction event; wherein the correlation evaluation rule includes the correlation analysis results on different evaluation dimensions and their corresponding correlation degree data.

4. The method according to claim 1, characterized in that: Before the step of determining the adverse drug reaction event description corresponding to the main key information in a preset medical review knowledge base as the target adverse drug reaction event description, the method further includes: Search the medical literature database for relevant literature on the adverse drug reaction events mentioned in the text of the medical report; Determine the adverse drug reaction event and the cause corresponding to the adverse drug reaction event in the relevant literature; The cause corresponding to the adverse drug reaction event is stored in a medical review knowledge base corresponding to the adverse drug reaction event as the cause of the adverse drug reaction.

5. A device for generating a medical review, applied to a first terminal, characterized in that: include: A medical report text receiving module, used for receiving a medical report text for a target drug sent by a second terminal; A key information acquisition module, used to extract key information from the medical report text; wherein the key information includes primary key information describing the adverse drug reaction event and secondary key information describing the cause of the adverse drug reaction event; wherein the secondary key information corresponds to different evaluation dimensions; A correlation analysis module, configured to perform correlation analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain correlation analysis results between the primary key information and the secondary key information in different evaluation dimensions; wherein the correlation analysis results include the existence of correlation or the absence of correlation; A correlation evaluation module, used to generate a correlation evaluation result between the target drug and the adverse drug reaction event based on the correlation analysis result and a preset correlation evaluation rule; A medical review report generating module, used for generating a medical review report according to the correlation evaluation result; The step of extracting key information from the medical report text by the key information acquisition module includes: Obtaining a preset medical review template; wherein the medical review template includes a plurality of slots; wherein the slots include attributes of information to be filled in; Extracting information corresponding to the attribute in the medical report text as key information; The step of extracting key information from the medical report text by the key information acquisition module further includes: adding attribute tags to the key information; wherein the attribute tags correspond to the attributes of the information to be filled in the slots of the medical review template; Accordingly, the step of generating a medical review report according to the relevance evaluation result by the medical review report generating module includes: Fill the key information and the relevance evaluation result into the relevance evaluation template according to the attribute label to obtain an evaluation conclusion report; Generate a medical review report based on the evaluation conclusion report; Wherein, after the step of extracting the key information in the medical report text by the key information acquisition module, the device is further used for: Acquire clinical research process information of the medical report text; wherein the clinical research process information includes at least one of the following: patient diagnosis process information, patient treatment process information, and patient treatment effect analysis process information; Filling the key information corresponding to the clinical research process information into the event summary template to obtain a drug adverse reaction event summary report; Accordingly, the step of generating a medical review report according to the relevance evaluation result by the medical review report generating module includes: Generate a medical review report based on the adverse drug reaction event summary report and the evaluation conclusion report; Wherein, the device is also used for: Fill the main key information and the secondary key information as well as the correlation analysis results between the main key information and the secondary key information into the correlation analysis template to obtain a correlation analysis report; Accordingly, the step of generating a medical review report according to the correlation evaluation result includes: generating a medical review report according to the adverse drug reaction event summary report, the correlation analysis report and the evaluation conclusion report; The step of performing association analysis on the primary key information and the secondary key information based on the evaluation dimension to obtain the analysis results of the association between the primary key information and the secondary key information on different evaluation dimensions includes: Determine the results of the correlation analysis between the primary key information and the secondary key information according to the context of the primary key information and the secondary key information in the text of the medical report; The step of the association analysis module performing association analysis on the primary key information and the secondary key information to obtain the association analysis result between the primary key information and the secondary key information further includes: Determining the adverse drug reaction event description corresponding to the main key information in a preset medical review knowledge base as the adverse drug reaction event description of the target drug; wherein the medical review knowledge base includes the target drug, the adverse reaction event of the target drug, and the cause of the adverse reaction event of the target drug; Calculating the similarity between the cause of the adverse drug reaction event corresponding to the target adverse drug reaction event description in the medical review knowledge base and the secondary key information; When the maximum value of the similarities is greater than or equal to a preset threshold, the result of the association analysis between the secondary key information and the primary key information is set as an association relationship; When the maximum value of the similarities is less than a preset threshold, the result of the association analysis between the secondary key information and the primary key information is set to indicate that there is no association relationship.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images