Case analysis method and system based on large language model

Abstract

The invention discloses a case analysis method and system based on a large language model, and belongs to the technical field of case analysis, and the method comprises the following steps: S1, patient information input, S2, examination item recommendation and determination, S3, preliminary diagnosis generation, S4, case screening and diagnosis and treatment suggestion generation, S5, diagnosis and treatment process monitoring and adjustment, and S6, case report generation and viewing. By means of the powerful knowledge integration and reasoning ability of a large language model, in combination with massive historical case data and professional medical database information, multi-dimensional diagnosis thought reference is provided for doctors, the misdiagnosis and missed diagnosis risks caused by personal experience limitation are reduced, the health indexes of patients can be monitored, and the diagnosis accuracy is improved. Personalized adjustment is performed on the diagnosis and treatment plan according to the disease change and the diagnosis and treatment condition of the patient; the medical database can dynamically expand and correct contents; diagnosis and treatment plans and suggestions can be given according to diagnosis and treatment passes of previously cured patients with similar conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of case analysis, and more specifically, to a case analysis method and system based on a large language model. Background Art

[0002] The foundation of case analysis methods lies in traditional medical knowledge. Medicine is an accumulative discipline, and thousands of years of medical practice have provided a rich theoretical basis for case analysis. For example, classic works such as Treatise on Febrile and Miscellaneous Diseases in traditional Chinese medicine and Corpus Hippocraticum in Western medicine contain descriptions of various disease symptoms, diagnoses, and treatment methods, and this knowledge is an important reference for early case analysis. Doctors classify and summarize the symptoms and signs of patients based on these classic theoretical knowledge, judge the disease type, and formulate treatment plans.

[0003] In existing case analysis technologies, there are often problems such as inaccurate diagnoses due to the limitations of doctors' personal experience, and the time-consuming and energy-consuming nature of doctors in retrieving cumbersome information and formulating plans, resulting in overly cumbersome course settings. At the same time, it is difficult to accurately detect patients' health indicators and it is difficult to make personalized adjustments to the diagnosis and treatment plan according to the changes in patients' conditions and diagnosis and treatment situations. Therefore, to solve the above problems, the existence of a case analysis method and system based on a large language model is crucial. Summary of the Invention

[0004] The main technical problem to be solved by the present invention is to provide a case analysis method and system based on a large language model, which can solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solution: A case analysis method based on a large language model, comprising the following steps:

[0006] S1. Patient information entry: On the front-end interface of the system, enter the patient's basic information in sequence according to the established fields, accurately record the patient's chief complaint, and write the current medical history completely according to the chronological order and the process of the disease development. At the same time, comprehensively sort out the patient's past medical history, conduct a comprehensive physical examination, enter the examination results, and upload the existing auxiliary examination results of the patient. After the entry is completed, the system automatically generates a unique patient ID and encrypts and stores the complete patient information in the local database;

[0007] S2. Recommended and determined examination items: The system performs desensitization processing on the stored patient information, only retains the medical key data, submits the desensitized patient information to the large language model through a pre-set interface for analysis and determination, and after the doctor confirms it is correct, clicks the submit button;

[0008] S3. Initial diagnosis generation: When the patient completes all the examinations prescribed by the doctor and the examination results are transmitted back to the case analysis system through the hospital information system, the system automatically merges and integrates the new examination results with the previously stored patient information. After desensitizing the integrated full patient information again, it is input into the large language model. Based on the latest data and combined with the medical knowledge graph, the large language model gives possible initial diagnoses. The system displays the initial diagnosis results on the doctor's end for the doctor to review and confirm. If the doctor has different opinions, they can make corrections according to the actual clinical situation and supplement detailed reasons.

[0009] S4. Case screening and treatment recommendation generation: With the initial diagnosis given by the large language model as the core, the system intelligently screens in the historical case database of the hospital's integrated information platform in combination with the patient's gender and age, and deeply analyzes each of the screened cases one by one to extract key treatment information. The system interacts with the medical database through a dedicated interface to obtain the latest treatments for related diseases. By comprehensively analyzing the above two aspects of information and combining the specific situation of the current patient, it uses built-in algorithms to generate a personalized treatment plan and at the same time gives detailed precautions, which are displayed on the doctor's end for their reference and optimization.

[0010] S5. Monitoring and adjustment during the treatment process: The system is connected to external medical devices or hardware in real time and continuously collects the patient's health data. For different patients, doctors can set personalized normal range values in the system in advance. Once it is found that a certain health data of the patient exceeds the preset normal range value, the system immediately sends an alarm to the doctor and details the time, value, and current basic information of the patient of the abnormal data. At the same time, the system automatically integrates all monitoring indicators and the previously stored patient information and inputs them into the large language model again. Based on the latest health data and the overall condition of the patient, the large language model gives treatment recommendations.

[0011] S6. Case report generation and viewing: When the patient completes the treatment and meets the discharge criteria, the system automatically starts the case report generation program, integrates and formats all the information of the patient during the entire treatment process in a standardized medical record format. The generated case report is stored in the system database. Doctors can view and download the case report at any time through the doctor's end for reviewing and summarizing treatment experience, conducting scientific research work, and carrying out medical quality control.

[0012] Furthermore, after accurately recording the patient's chief complaint in S1, based on the patient's exact words, the core discomfort symptoms that prompted them to seek medical treatment and the duration from the first occurrence to the present are sorted out. The medical history covers the initial manifestations of the symptoms, factors for aggravation and remission, accompanying symptoms, whether self-medication has been taken and the medication effect.

[0013] Further, when comprehensively sorting out the patient's past medical history in S1, regarding the personal history, understand the patient's smoking history, drinking history, occupation and occupational exposure, travel history, and for the reproductive history, ask relevant information according to the patient's gender and age, and carefully inquire about the family history, focusing on the incidence of genetic diseases and mental diseases in the family, and draw a simple family tree to assist in recording.

[0014] Further, in S2, the large language model, based on the medical knowledge it has trained and its understanding of similar cases, quickly analyzes and generates a list of examination items that may need to be completed. After the doctor submits in S2, the system automatically sends the application for the prescribed examination items to the hospital's inspection and examination management system, and at the same time generates an examination appointment form for the patient to view, and tracks the examination progress in real time, waiting for the examination results to be reported.

[0015] Further, the doctor confirmation step in S2 is as follows:

[0016] S2.1. After the doctor receives the examination item suggestions feedback by the large language model on the system's interaction interface, combine their own professional knowledge, clinical experience, and understanding of the patient's actual situation to review each item one by one;

[0017] S2.2. If it is recognized that the item is reasonable and necessary, directly check and confirm;

[0018] S2.3. If it is considered that the item is partially reasonable but needs to be adjusted, edit and modify it on the basis of the original item;

[0019] S2.4. If it is determined that the item is not applicable to the current patient, delete it.

[0020] Further, the screening conditions set in S4 are that the preliminary diagnoses are the same, the genders are the same, and the ages are within the range of ±5 years, and a batch of cured and discharged cases with reference value are screened out.

[0021] Further, during the diagnosis and treatment process in S5, if the doctor finds that the patient's condition has new changes and there are new examination results reported, the doctor can input this information into the system through the doctor's terminal at any time. The system quickly integrates all the information of this patient, re-enters it into the large language model, and the large language model gives suggestions on the examination items that need to be further improved, or continues the current diagnosis and treatment plan, or modifies the diagnosis and treatment items. The doctor adjusts the diagnosis and treatment plan accordingly.

[0022] According to another aspect of the present invention, there is provided a case analysis system based on a large language model, including an information input module, an information storage module, an information desensitization module, an information interaction module, an examination item management module, a case screening module, a case analysis module, a health data monitoring module, a diagnosis process optimization module, and a case report generation and viewing module;

[0023] The health data monitoring module includes an external connection module, a data modification module, and a data comparison module;

[0024] The external connection module is used to establish a real-time connection with external medical devices or hardware and continuously collect patients' health data. The data modification module is used to allow doctors to individually modify the normal range value of a certain health data in the system according to individual differences for different patients. The data comparison module is used to compare the real-time monitoring data with the preset normal range value at regular intervals. Once data beyond the normal range value is found, it immediately notifies the doctor through the system, and at the same time integrates and records the abnormal data and relevant patient information to provide a basis for subsequent condition analysis.

[0025] Furthermore, the information input module is responsible for receiving all-round information of patients. The information storage module is used to generate a unique patient ID according to a specific algorithm and store the complete patient information in the local database at the same time;

[0026] The information desensitization module is used to remove sensitive information that may disclose the patient's identity using a desensitization algorithm before submitting the patient information to the large language model, and only retain the key medical data related to disease diagnosis and treatment;

[0027] The information interaction module is used to establish a stable interaction interface with the large language model, responsible for sending the desensitized patient information to the large language model and receiving various analysis results returned by the model;

[0028] The examination item management module is used to receive the examination items that may need to be improved recommended by the large language model, display them in the doctor's operation interface in the form of a list, and automatically dock with the hospital inspection and examination system to complete the issuance of examination applications and track the examination progress;

[0029] The case screening module is used to be deeply integrated with the hospital integrated information platform and quickly retrieve and match cases in the massive historical case database according to the set screening conditions;

[0030] The case analysis module is used to conduct a detailed analysis of the screened case information and the diagnosis and treatment process, extract valuable information such as key diagnosis and treatment nodes, the basis for the selection and adjustment of treatment plans, and the rehabilitation period. At the same time, it interacts with the medical database through a dedicated interface to obtain auxiliary diagnosis and treatment information, and combines the current patient situation to comprehensively generate personalized diagnosis and treatment suggestions;

[0031] The diagnosis process optimization module is used to receive information such as changes in the patient's condition and new examination results input by the doctor from the doctor's end at any time during the diagnosis and treatment process, integrate these dynamic information with all the patient information stored in the system in real time, and submit it to the large language model again to obtain the examination suggestions that need to be further improved given by the model;

[0032] The case report generation and viewing module is used to integrate and typeset all the information of the patient during the entire diagnosis and treatment process in accordance with the standardized medical record format to generate a complete and detailed case report.

[0033] The beneficial effects of a case analysis method and system based on a large language model of the present invention are as follows:

[0034] With the powerful knowledge integration and reasoning capabilities of the large language model, combined with a large amount of historical case data and information in professional medical databases, it provides doctors with multi-dimensional diagnostic thinking references, reduces the risk of misdiagnosis and missed diagnosis caused by personal experience limitations. Especially for rare diseases and complex and difficult diseases, it can quickly associate similar cases to assist in accurate diagnosis, thereby improving the accuracy of diagnosis;

[0035] From the intelligent recommendation of examination items, the personalized generation of treatment plans to the dynamic adjustment suggestions during the diagnosis and treatment process, the system runs through the entire diagnosis and treatment process, reducing the time and energy consumed by doctors in retrieving cumbersome information and formulating plans, making the diagnosis and treatment process more efficient and smooth, shortening the waiting time for patients to seek medical treatment, improving the overall efficiency of medical services, and thus optimizing the diagnosis and treatment process;

[0036] By continuously monitoring the patient's health data and flexibly adjusting the normal range values in combination with individual differences, the system can keenly capture the subtle changes in the patient's condition, give early warnings in a timely manner, provide strong support for doctors to implement precise interventions, truly achieve individualized treatment that varies from person to person and from disease to disease, improve the treatment effect, improve the patient's prognosis, and thus achieve precision medicine;

[0037] Through the standardized generation and convenient viewing of case reports, it is not only convenient for doctors to summarize and review individual cases. In the long run, it has formed a rich clinical case knowledge base within the hospital, providing a large number of real and detailed first-hand materials for the training of young doctors and medical research, promoting the inheritance and sharing of medical knowledge within the team, and promoting the development of disciplines, thereby promoting the inheritance and accumulation of medical knowledge. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will be further described in detail below with reference to the drawings and specific implementation methods.

[0039] Figure 1 It is a schematic flowchart of the method of a case analysis method and system based on a large language model of the present invention;

[0040] Figure 2 It is a schematic diagram of the system module structure of a case analysis method and system based on a large language model of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] To make the technical solutions of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0042] Example 1

[0043] As shown in the figure, according to one aspect of the present invention, a technical solution is provided: a case analysis method based on a large language model, including the following steps:

[0044] Step 1: Patient information entry: On the front-end interface of the system, input the patient's basic information in sequence according to the established fields. At the same time, accurately record the patient's chief complaint, and write the current medical history completely in chronological order and according to the process of the disease development. At the same time, comprehensively sort out the patient's past medical history, and conduct a comprehensive and systematic physical examination. Enter the examination results, and upload the existing auxiliary examination results of the patient. After the entry is completed, the system automatically generates a unique patient ID, and encrypts and stores the complete patient information in the local database. After accurately recording the patient's chief complaint, based on the patient's exact words, sort out the core discomfort symptoms that prompted them to seek medical treatment and the duration from the first appearance to the current. The medical history covers the initial manifestations of symptoms, factors for aggravation and remission, accompanying symptoms, whether self-medication has been taken and the medication effect. When comprehensively sorting out the patient's past medical history, for the personal history, understand the patient's smoking history, drinking history, occupation and occupational exposure, travel history, and for the reproductive history, ask relevant information according to the patient's gender and age, and carefully ask about the family history, focusing on the incidence of genetic diseases and mental diseases in the family, and draw a simple family tree to assist in recording.

[0045] Step 2: Recommended and determined examination items: The system performs desensitization processing on the stored patient information, only retaining the key medical data, and submits the desensitized patient information to the large language model through a pre-set interface for analysis and determination. After the doctor confirms that it is correct, click the submit button. Based on the medical knowledge it has trained and the understanding of similar cases, the large language model quickly analyzes and generates a list of examination items that may need to be improved. After the doctor submits, the system automatically sends the application for the prescribed examination items to the hospital's inspection and examination management system, and at the same time generates an examination appointment form for the patient to view, and tracks the examination progress in real time, waiting for the return of the examination results. The doctor's confirmation steps are as follows:

[0046] S2.1. After receiving the examination item suggestions fed back by the large language model on the interactive interface of the system, the doctor reviews each item one by one in combination with their own professional knowledge, clinical experience and understanding of the actual situation of the patient;

[0047] S2.2. If it is recognized that the item is reasonable and necessary, directly check and confirm;

[0048] S2.3. If it is considered that the item is partially reasonable but needs to be adjusted, edit and modify it on the basis of the original item;

[0049] S2.4. If it is determined that the item is not applicable to the current patient, delete it.

[0050] Step 3. Generation of preliminary diagnosis: When the patient completes all the examinations prescribed by the doctor and the examination results are transmitted back to the case analysis system through the hospital information system, the system automatically merges and integrates the new examination results with the previously stored patient information. After desensitizing the integrated full patient information again, it is input into the large language model. Based on the latest data and combined with the medical knowledge graph, the large language model gives possible preliminary diagnoses. The system displays the preliminary diagnosis results on the doctor's side for the doctor to review and confirm. If the doctor has different opinions, they can make corrections according to the clinical actual situation and supplement detailed reasons.

[0051] Step 4. Case screening and generation of treatment suggestions: With the preliminary diagnosis given by the large language model as the core, the system combines the patient's gender and age to conduct intelligent screening in the historical case database of the hospital's integrated information platform, and conducts in-depth analysis on each of the screened cases to extract key treatment information. The system interacts with the medical database through a dedicated interface to obtain the latest treatments for related diseases. By comprehensively analyzing the above two aspects of information and combining the specific situation of the current patient, it uses built-in algorithms to generate personalized treatment plans and at the same time gives detailed precautions, which are displayed on the doctor's side for their reference and optimization. The screening conditions are set as the same preliminary diagnosis, the same gender, and the age within the range of ±5 years, and a batch of cured and discharged cases with reference value are screened out.

[0052] 7. A case analysis method based on a large language model according to claim 1.

[0053] Step 5. Monitoring and adjustment during the treatment process: The system is connected to external medical devices or hardware in real time to continuously collect the patient's health data. For different patients, doctors can set personalized normal range values in the system in advance. Once it is found that a certain health data of the patient exceeds the preset normal range value, the system immediately sends an alarm to the doctor and details the time, value of the abnormal data, and the current basic information of the patient. At the same time, the system automatically integrates all monitoring indicators and the previously stored patient information and inputs it into the large language model again. Based on the latest health data and the overall condition of the patient, the large language model gives treatment suggestions. During the treatment process, if the doctor finds new changes in the patient's condition and new examination results are reported, they can input this information into the system through the doctor's side at any time. The system quickly integrates all the information of this patient and inputs it into the large language model again. The large language model gives suggestions on the examinations that need to be further improved, continuing the current treatment plan, or modifying the treatment items. The doctor adjusts the treatment plan in a timely manner based on this.

[0054] Step 6: Case Report Generation and Viewing: When the patient completes the diagnosis and treatment and meets the discharge criteria, the system automatically starts the case report generation program, integrates and formats all the information of the patient during the entire diagnosis and treatment process in a standardized medical record format. The generated case report is stored in the system database, and doctors can view and download the case report at any time through the doctor's terminal for reviewing and summarizing diagnosis and treatment experience, conducting scientific research, and performing medical quality control.

[0055] The method also sets up a case analysis system based on a large language model. The system includes an information input module, an information storage module, an information desensitization module, an information interaction module, an examination item management module, a case screening module, a case analysis module, a health data monitoring module, a diagnosis process optimization module, and a case report generation and viewing module;

[0056] The health data monitoring module includes an external connection module, a data modification module, and a data comparison module;

[0057] The external connection module is used to establish a real-time connection with external medical devices or hardware and continuously collect the patient's health data. The data modification module is used to allow doctors to separately modify the normal range value of a certain health data in the system according to individual differences for different patients. The data comparison module is used to compare the real-time monitoring data with the preset normal range value at regular intervals. Once data beyond the normal range value is found, the doctor is immediately notified through the system, and at the same time, the abnormal data and relevant patient information are integrated and recorded to provide a basis for subsequent condition analysis;

[0058] The information input module is responsible for receiving all-round information of the patient. The information storage module is used to generate a unique patient ID according to a specific algorithm and store the complete patient information in the local database;

[0059] The information desensitization module is used to remove sensitive information that may disclose the patient's identity using a desensitization algorithm before submitting the patient information to the large language model, and only retain the key medical data related to disease diagnosis and treatment;

[0060] The information interaction module is used to establish a stable interaction interface with the large language model, responsible for sending the desensitized patient information to the large language model and receiving various analysis results returned by the model;

[0061] The examination item management module is used to receive the examination items that may need to be improved recommended by the large language model, display them in a list form on the doctor's operation interface, and automatically dock with the hospital inspection and examination system to issue examination applications and track the examination progress;

[0062] The case screening module is used to be deeply integrated with the hospital integrated information platform and quickly retrieve and match cases in the massive historical case database according to the set screening conditions;

[0063] The case analysis module is used to conduct a detailed analysis of the selected case information and the diagnosis and treatment process, extract valuable information such as key diagnosis and treatment nodes, the basis for the selection and adjustment of treatment plans, and the rehabilitation period. At the same time, it interacts with the medical database through a dedicated interface to obtain auxiliary diagnosis and treatment information, and combines the current patient situation to comprehensively generate personalized diagnosis and treatment suggestions;

[0064] The diagnosis process optimization module is used to receive information such as changes in the patient's condition and new examination results input by the doctor from the doctor side at any time during the diagnosis and treatment process, integrate these dynamic information with all the patient's data stored in the system in real time, and submit them to the large language model again to obtain examination suggestions that need to be further improved given by the model;

[0065] The case report generation and viewing module is used to integrate and typeset all the information of the patient during the entire diagnosis and treatment process in accordance with the standard medical record format to generate a complete and detailed case report.

[0066] The above embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the appended claims.

Claims

1. A case analysis method based on a large language model, characterized in that: The following steps are involved: S1. Patient information entry: On the front-end interface of the system, enter the patient's basic information in the specified fields in sequence, accurately record the patient's chief complaint, and write a complete medical history in chronological order and according to the course of the disease. At the same time, comprehensively sort out the patient's past history, conduct a comprehensive and systematic physical examination, enter the examination results, and upload the patient's existing auxiliary examination results. After the entry is completed, the system automatically generates a unique patient ID and encrypts the complete patient information and stores it in the local database; S2. Recommendation and confirmation of examination items: The system desensitizes the stored patient information and only retains key medical data. The desensitized patient information is submitted to the large language model through a pre-set interface for analysis and confirmation. After the doctor confirms that it is correct, he clicks the Submit button; S3. Preliminary diagnosis generation: When the patient completes the various examinations prescribed by the doctor, the examination results are transmitted back to the case analysis system through the hospital information system. The system automatically merges the new examination results with the previously stored patient information, and then desensitizes the integrated patient information and inputs it into the large language model. The large language model gives a possible preliminary diagnosis based on the latest data and the medical knowledge graph. The system displays the preliminary diagnosis results on the doctor's side, and the doctor reviews and confirms them. If the doctor has different opinions, he can make corrections based on the actual clinical situation and provide detailed reasons. S4. Case screening and diagnosis and treatment recommendation generation: The system takes the preliminary diagnosis given by the large language model as the core, combines the patient's gender and age, and intelligently screens the historical case database of the hospital's integrated information platform. It also conducts in-depth analysis of the screened cases one by one to extract key diagnosis and treatment information. The system interacts with the medical database through a dedicated interface to obtain the latest diagnosis and treatment of related diseases, comprehensively analyzes the above two aspects of information, combines the specific situation of the current patient, and uses the built-in algorithm to generate a personalized diagnosis and treatment plan. At the same time, it gives detailed precautions and displays them on the doctor's side for reference and optimization; S5. Monitoring and adjustment of the diagnosis and treatment process: The system is connected to external medical equipment or hardware in real time to continuously collect the patient's health data. For different patients, doctors can set personalized normal range values ​​in the system in advance. Once it is found that a certain health data of the patient exceeds the preset normal range value, the system will immediately send an alarm to the doctor and display the time, value and current basic information of the abnormal data in detail. At the same time, the system automatically integrates all monitoring indicators and previously stored patient information, and inputs them into the large language model again. The large language model gives diagnosis and treatment suggestions based on the latest health data and the patient's overall condition; S6. Case report generation and review: When the patient completes diagnosis and treatment and meets the discharge criteria, the system automatically starts the case report generation program, and integrates and formats all the patient's information throughout the diagnosis and treatment process in a standardized medical record format. The generated case report is stored in the system database. Doctors can view and download the case report at any time through the doctor's terminal to review and summarize diagnosis and treatment experience, conduct scientific research and perform medical quality control.

2. The case analysis method based on a large language model according to claim 1, characterized in that: After accurately recording the patient's chief complaint in S1, the core discomfort symptoms that prompted the patient to seek medical treatment and their duration from the first appearance to the present are sorted out based on the patient's original words. The medical history covers the onset of symptoms, aggravating and relieving factors, accompanying symptoms, whether the patient takes medication on his own, and the effects of medication.

3. The case analysis method based on a large language model according to claim 1, characterized in that: When comprehensively sorting out the patient's past history in S1, the patient's smoking history, drinking history, occupation and occupational exposure, and travel history are understood for personal history. For marital history, relevant information is asked based on the patient's gender and age, and family history is carefully inquired, with a focus on the incidence of genetic diseases and mental illnesses in the family, and a simple family tree is drawn to assist in recording.

4. The case analysis method based on a large language model according to claim 1, characterized in that: The large language model in S2 quickly analyzes and generates a list of examination items that may need to be improved based on its trained medical knowledge and understanding of similar cases. After the doctor in S2 submits the application, the system automatically sends the application for the issued examination items to the hospital's inspection and examination management system, and generates an examination appointment form for the patient to review, and tracks the examination progress in real time, waiting for the examination results to be reported.

5. The case analysis method based on a large language model according to claim 1, characterized in that: The doctor confirmation step in S2 is: S2.

1. After receiving the examination item suggestions from the large language model on the interactive interface of the system, the doctor will review each item one by one based on his or her own professional knowledge, clinical experience and understanding of the patient's actual situation; S2.

2. If the approved project is reasonable and necessary, just check the confirmation box; S2.

3. If it is considered that part of the project is reasonable but needs to be adjusted, edit and modify it based on the original project; S2.

4. If the item is judged not to be applicable to the current patient, it shall be deleted.

6. The case analysis method based on a large language model according to claim 1, characterized in that: The screening conditions set in S4 are the same preliminary diagnosis, the same gender, and the age within the range of ±5 years, so as to screen out a group of cured and discharged cases with reference value.

7. The case analysis method based on a large language model according to claim 1, characterized in that: In S5, during the diagnosis and treatment process, if the doctor finds new changes in the patient's condition and new test results are reported, the doctor can input this information into the system at any time through the doctor's end. The system quickly integrates all the patient's information and re-enters the large language model. The large language model gives suggestions for further examinations that need to be improved or for continuing the current diagnosis and treatment plan or modifying the diagnosis and treatment items. The doctor adjusts the diagnosis and treatment plan in a timely manner based on this.

8. A case analysis system based on a large language model, comprising the case analysis method based on a large language model according to any one of claims 1 to 7, characterized in that: It includes information input module, information storage module, information desensitization module, information interaction module, examination project management module, case screening module, case analysis module, health data monitoring module, diagnosis process optimization module and case report generation and viewing module; The health data monitoring module includes an external connection module, a data modification module and a data comparison module; The external connection module is used to establish a real-time connection with external medical equipment or hardware to continuously collect the patient's health data. The data modification module is used to allow doctors to modify the normal range value of a certain health data in the system according to individual differences for different patients. The data comparison module is used to compare the real-time monitoring data with the preset normal range value at regular intervals. Once data exceeding the normal range value is found, the doctor is immediately notified through the system. At the same time, the abnormal data and related patient information are integrated and recorded to provide a basis for subsequent disease analysis.

9. The case analysis system based on a large language model according to claim 8, characterized in that: The information input module is responsible for receiving all-round information of the patient, and the information storage module is used to generate a unique patient ID according to a specific algorithm, and at the same time, store the complete patient information in a local database; The information desensitization module is used to remove sensitive information that may reveal the patient's identity by using a desensitization algorithm before submitting the patient information to the large language model, and only retain key medical data related to disease diagnosis and treatment; The information interaction module is used to establish a stable interactive interface with the large language model, is responsible for sending the desensitized patient information to the large language model, and receives various analysis results returned by the model; The inspection project management module is used to receive the inspection items that may need to be improved as recommended by the large language model, and display them in the form of a list on the doctor's operation interface, and automatically connect with the hospital's inspection and examination system to complete the issuance of inspection applications and track the progress of inspections; The case screening module is used to deeply integrate with the hospital's integrated information platform to quickly retrieve matching cases in a massive historical case database based on set screening conditions; The case analysis module is used to analyze the screened case information and diagnosis and treatment process in detail, extract valuable information such as key diagnosis and treatment nodes, selection and adjustment basis of treatment plans, rehabilitation cycle, etc. At the same time, it interacts with the medical database through a dedicated interface to obtain auxiliary diagnosis and treatment information, and comprehensively generates personalized diagnosis and treatment suggestions based on the current patient situation; The diagnostic process optimization module is used to receive information such as changes in the patient's condition and new examination results input by the doctor from the doctor's end at any time during the diagnosis and treatment process, integrate these dynamic information with all the patient information stored in the system in real time, and submit them to the large language model again to obtain the examination suggestions given by the model that need to be further improved; The case report generation and viewing module is used to integrate and typeset all the information of the patient throughout the entire diagnosis and treatment process in accordance with the standardized medical record format to generate a complete and detailed case report.

Images