AI-based medical guidance system for monitoring and feedback of patient disease information

Abstract

The present invention discloses a method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system, comprising the following steps: S1, storing medical record data in a first database; S2, obtaining multiple disease labels, multiple disease symptom data and multiple treatment suggestion data; S3, when the similarity is greater than a similarity threshold, the first control module recommends the disease label corresponding to the disease symptom data and the treatment suggestion data corresponding to the disease label to a display module; using the above method, collecting and storing past records of the same type of diseases; classifying and arranging procedural problems in these records; when a doctor fills in relevant disease symptoms, automatically retrieving relevant records and providing corresponding suggestions to the user, which can also prevent the doctor from forgetting some precautions.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method for monitoring and providing feedback on patient disease information in an AI-based diagnosis guidance system. Background Art

[0002] In the medical field, with the continuous development of information technology, the application of electronic medical records and health information systems is becoming more and more widespread. The experience accumulated by doctors in the diagnosis and treatment process has important reference value for subsequent patients. However, at present, when most medical institutions deal with the same type of cases, doctors' advice and experience are often limited to individual communication and records, lacking systematic organization and utilization, which leads to the failure to effectively share valuable knowledge resources.

[0003] In addition, when it comes to procedural issues, that is, those that are not clinical decisions but have an important impact on the medical process, such as appointment processes, medication guidelines, follow-up appointment arrangements, etc., the existing system usually does not have a unified and efficient way to manage and provide feedback to doctors or patients who need help. When patients see a doctor, they can only make judgments based on the doctor's medical experience. However, since doctors cannot remember the detailed precautions for each condition, it will have a greater impact on the patient's treatment experience and effect. Summary of the invention

[0004] In view of this, an object of the present invention is to provide an AI-based medical guidance system for monitoring and providing feedback on patient disease information, which feedback method can automatically recommend treatment suggestions.

[0005] The technical solution adopted by the present invention to solve its technical problem is to provide a method for monitoring and feedback of patient disease information in a diagnosis guidance system based on AI, comprising: a server and a plurality of user terminals connected to the server by signal, each user terminal comprising a display module, an input module, a first control module and a first database, and the server comprising a second control module and a second database;

[0006] The feedback method includes the following steps:

[0007] S1. The user inputs the patient's medical record through the input module to form medical record data, and the first control module stores the medical record data into the first database;

[0008] S2. The first control module is preset with a standard upload interval. The first control module regularly transmits all medical record data in the first database to the second database according to the standard upload interval. The second control module processes all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data, each disease label corresponds to at least one disease symptom data, and each disease label corresponds to at least one treatment suggestion data;

[0009] The second control module is preset with a standard update interval, and the second control module regularly transmits all disease labels, disease symptom data and treatment suggestion data to the first database according to the standard update interval;

[0010] S3. While the user inputs the patient's medical record through the input module, the first control module processes the patient's medical record through natural language processing technology to extract the medical symptom data. The first control module calculates the similarity between the medical symptom data and all disease symptom data under the same disease label. The first control module has a preset similarity threshold. When the similarity is greater than the similarity threshold, the first control module recommends the disease label corresponding to the disease symptom data and the treatment recommendation data corresponding to the disease label to the display module.

[0011] As a further improvement of the present invention, the specific steps of the second control module in step S2 processing all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data include:

[0012] S21, the second control module pre-processes all medical record data to obtain pre-processed medical record data;

[0013] S22, performing entity recognition and labeling on the medical record preprocessing data, thereby obtaining medical record entity data;

[0014] S23. Extract multiple disease labels, multiple disease symptom data, and multiple treatment suggestion data based on the medical record entity data.

[0015] As a further improvement of the present invention, the specific steps of preprocessing in step S21 include:

[0016] S211, remove format symbols in medical record data;

[0017] S212, segmenting the continuous text into multiple medical terms;

[0018] S213. Unify medical terms for the same concept.

[0019] As a further improvement of the present invention, the specific steps of performing entity recognition and labeling on the medical record preprocessing data in step S22 to obtain medical record entity data are: substituting the medical record preprocessing data into the medical NER model for recognition and labeling to obtain multiple medical entities, and integrating all medical entities to form medical record entity data.

[0020] As a further improvement of the present invention, the multiple medical entities are disease name, symptom description, organ location, drug name, follow-up visit time and examination items.

[0021] As a further improvement of the present invention, the specific steps of extracting multiple disease labels, multiple disease symptom data and multiple treatment suggestion data according to the medical record entity data in step S23 include:

[0022] S231. Classify and label the medical entities identified as disease names according to the ICD International Classification of Diseases, thereby obtaining multiple disease labels;

[0023] S232, classifying and labeling the medical entities identified as symptom descriptions and organ locations according to a hierarchical clustering algorithm, thereby obtaining a plurality of disease symptom data;

[0024] S233. The medical entities identified as drug names, follow-up visit times, and examination items are recorded as treatment recommendation data.

[0025] As a further improvement of the present invention, the specific steps of processing in step S3 include:

[0026] S31. Remove formatting symbols from patient medical records;

[0027] S32, segmenting the continuous text into multiple medical terms;

[0028] S33. Unify medical terms for the same concept.

[0029] As a further improvement of the present invention, the first control module is preset with a bag-of-words model;

[0030] Step S2.5 is also included after step S2: the first control module substitutes all disease symptom data into the bag-of-words model to convert the disease symptom data into a standard symptom feature vector;

[0031] In step S3, the first control module performs similarity calculation between the medical symptom data and all disease symptom data corresponding to the same disease label, including:

[0032] S34, the first control module substitutes the medical symptom data into the bag-of-words model, and converts all medical terms in the medical symptom data into medical feature vectors;

[0033] S35. Calculate the cosine value of the angle between the medical consultation feature vector and the standard symptom feature vector.

[0034] As a further improvement of the present invention, the multiple disease symptom data are fever, cough, headache, abdominal pain, chest tightness, fatigue, loss of appetite, nausea and vomiting, rash and joint pain.

[0035] As a further improvement of the present invention, the plurality of treatment suggestion data are respectively taking one ibuprofen sustained-release tablet after meals twice a day, acupuncture treatment three times a week and reducing salt intake.

[0036] The beneficial effects of the present invention are at least that the platform can: collect and store past records of similar diseases; classify and organize procedural problems in these records; automatically retrieve relevant records and provide corresponding suggestions to users when doctors fill in relevant disease symptoms, which can also prevent doctors from forgetting some precautions; use NLP technology to extract medical symptom data from medical record texts, and match them with standardized disease labels to ensure the consistency and accuracy of symptom descriptions; and through the continuous collection and processing of large amounts of medical record data, it helps to discover new disease association patterns. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0038] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0039] Reference Figure 1 , the present invention proposes a method for monitoring and feedback of patient disease information in a diagnosis guidance system based on AI, comprising: a server and a plurality of user terminals connected to the server by signal, each user terminal comprising a display module, an input module, a first control module and a first database, and the server comprising a second control module and a second database;

[0040] The feedback method includes the following steps:

[0041] S1. The user inputs the patient's medical record through the input module to form medical record data, and the first control module stores the medical record data into the first database;

[0042] S2. The first control module is preset with a standard upload interval. The first control module regularly transmits all medical record data in the first database to the second database according to the standard upload interval. The second control module processes all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data, each disease label corresponds to at least one disease symptom data, and each disease label corresponds to at least one treatment suggestion data;

[0043] The second control module is preset with a standard update interval, and the second control module regularly transmits all disease labels, disease symptom data and treatment suggestion data to the first database according to the standard update interval;

[0044] S3. While the user inputs the patient's medical record through the input module, the first control module processes the patient's medical record through natural language processing technology to extract the medical symptom data. The first control module calculates the similarity between the medical symptom data and all disease symptom data under the same disease label. The first control module has a preset similarity threshold. When the similarity is greater than the similarity threshold, the first control module recommends the disease label corresponding to the disease symptom data and the treatment recommendation data corresponding to the disease label to the display module.

[0045] Collect and store past records of similar diseases; classify and organize procedural issues in these records; automatically retrieve relevant records and provide users with corresponding suggestions when doctors fill in relevant disease symptoms, which can also prevent doctors from forgetting some precautions; use NLP technology to extract medical symptom data from medical record texts and match them with standardized disease labels to ensure the consistency and accuracy of symptom descriptions; through the continuous collection and processing of large amounts of medical record data, it helps to discover new disease association patterns.

[0046] As a further improvement of the present invention, the specific steps of the second control module in step S2 processing all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data include:

[0047] S21, the second control module pre-processes all medical record data to obtain pre-processed medical record data;

[0048] S22, performing entity recognition and labeling on the medical record preprocessing data, thereby obtaining medical record entity data;

[0049] S23. Extract multiple disease labels, multiple disease symptom data, and multiple treatment suggestion data based on the medical record entity data.

[0050] Ensure that all data processing follows unified standards to reduce the risk of misdiagnosis due to data errors; for primary medical institutions or remote areas, it can be used as an auxiliary tool for diagnosis and treatment, helping local doctors better understand the condition and make reasonable treatment choices, alleviating the problem of insufficient medical resources.

[0051] As a further improvement of the present invention, the specific steps of preprocessing in step S21 include:

[0052] S211, remove format symbols in medical record data;

[0053] S212, segmenting the continuous text into multiple medical terms;

[0054] S213. Unify medical terms for the same concept.

[0055] As a further improvement of the present invention, the specific steps of performing entity recognition and labeling on the medical record preprocessing data in step S22 to obtain medical record entity data are: substituting the medical record preprocessing data into the medical NER model for recognition and labeling to obtain multiple medical entities, and integrating all medical entities to form medical record entity data.

[0056] As a further improvement of the present invention, the multiple medical entities are disease name, symptom description, organ location, drug name, follow-up visit time and examination items.

[0057] As a further improvement of the present invention, the specific steps of extracting multiple disease labels, multiple disease symptom data and multiple treatment suggestion data according to the medical record entity data in step S23 include:

[0058] S231. Classify and label the medical entities identified as disease names according to the ICD International Classification of Diseases, thereby obtaining multiple disease labels;

[0059] S232, classifying and labeling the medical entities identified as symptom descriptions and organ locations according to a hierarchical clustering algorithm, thereby obtaining a plurality of disease symptom data;

[0060] S233. The medical entities identified as drug names, follow-up visit times, and examination items are recorded as treatment recommendation data.

[0061] The ICD International Classification of Diseases system is used to classify and label disease names to ensure the consistency and standardization of disease labels, reduce confusion caused by different descriptions of the same disease among different doctors or medical institutions, and facilitate cross-regional and cross-institutional data sharing and communication; disease labels based on ICD classification can more accurately reflect the type and severity of the disease; the application of hierarchical clustering algorithms makes the description of symptoms and the classification of organ locations more detailed.

[0062] As a further improvement of the present invention, the specific steps of processing in step S3 include:

[0063] S31. Remove formatting symbols from patient medical records;

[0064] S32, segmenting the continuous text into multiple medical terms;

[0065] S33. Unify medical terms for the same concept.

[0066] As a further improvement of the present invention, the first control module is preset with a bag-of-words model;

[0067] Step S2.5 is also included after step S2: the first control module substitutes all disease symptom data into the bag-of-words model to convert the disease symptom data into a standard symptom feature vector;

[0068] In step S3, the first control module performs similarity calculation between the medical symptom data and all disease symptom data corresponding to the same disease label, including:

[0069] S34, the first control module substitutes the medical symptom data into the bag-of-words model, and converts all medical terms in the medical symptom data into medical feature vectors;

[0070] S35. Calculate the cosine value of the angle between the medical consultation feature vector and the standard symptom feature vector.

[0071] As a further improvement of the present invention, the multiple disease symptom data are fever, cough, headache, abdominal pain, chest tightness, fatigue, loss of appetite, nausea and vomiting, rash and joint pain.

[0072] As a further improvement of the present invention, the plurality of treatment suggestion data are respectively taking one ibuprofen sustained-release tablet after meals twice a day, acupuncture treatment three times a week and reducing salt intake.

Claims

1. A method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system, characterized in that: include: A server and a plurality of user terminals connected to the server by signal, each of which comprises a display module, an input module, a first control module and a first database, and the server comprises a second control module and a second database; The feedback method includes the following steps: S1. The user inputs the patient's medical record through the input module to form medical record data, and the first control module stores the medical record data into the first database; S2. The first control module is preset with a standard upload interval. The first control module regularly transmits all medical record data in the first database to the second database according to the standard upload interval. The second control module processes all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data, each disease label corresponds to at least one disease symptom data, and each disease label corresponds to at least one treatment suggestion data; The second control module is preset with a standard update interval, and the second control module regularly transmits all disease labels, disease symptom data and treatment suggestion data to the first database according to the standard update interval; S3. While the user inputs the patient's medical record through the input module, the first control module processes the patient's medical record through natural language processing technology to extract the medical symptom data. The first control module calculates the similarity between the medical symptom data and all disease symptom data under the same disease label. The first control module has a preset similarity threshold. When the similarity is greater than the similarity threshold, the first control module recommends the disease label corresponding to the disease symptom data and the treatment recommendation data corresponding to the disease label to the display module.

2. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 1, characterized in that: In step S2, the second control module processes all medical record data by natural language processing technology to obtain multiple disease labels, multiple disease symptom data and multiple treatment suggestion data. The specific steps include: S21, the second control module pre-processes all medical record data to obtain pre-processed medical record data; S22, performing entity recognition and labeling on the medical record preprocessing data, thereby obtaining medical record entity data; S23. Extract multiple disease labels, multiple disease symptom data, and multiple treatment suggestion data based on the medical record entity data.

3. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 2, characterized in that: The specific steps of preprocessing in step S21 include: S211, remove format symbols in medical record data; S212, segmenting the continuous text into multiple medical terms; S213. Unify medical terms for the same concept.

4. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 2, characterized in that: The specific steps of performing entity recognition and labeling on the medical record preprocessing data in step S22 to obtain medical record entity data are: substituting the medical record preprocessing data into the medical NER model to identify and label multiple medical entities, and integrating all medical entities to form medical record entity data.

5. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 4, characterized in that: The multiple medical entities are disease name, symptom description, organ location, drug name, follow-up visit time and examination items.

6. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 5, characterized in that: The specific steps of extracting multiple disease labels, multiple disease symptom data and multiple treatment suggestion data according to the medical record entity data in step S23 include: S231. Classify and label the medical entities identified as disease names according to the ICD International Classification of Diseases, thereby obtaining multiple disease labels; S232, classifying and labeling the medical entities identified as symptom descriptions and organ locations according to a hierarchical clustering algorithm, thereby obtaining a plurality of disease symptom data; S233. The medical entities identified as drug names, follow-up visit times, and examination items are recorded as treatment recommendation data.

7. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 1, characterized in that: The specific steps of processing in step S3 include: S31. Remove formatting symbols from patient medical records; S32, segmenting the continuous text into multiple medical terms; S33. Unify medical terms for the same concept.

8. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 7, characterized in that: The first control module is preset with a bag-of-words model; Step S2.5 is also included after step S2: the first control module substitutes all disease symptom data into the bag-of-words model to convert the disease symptom data into a standard symptom feature vector; In step S3, the first control module performs similarity calculation between the medical symptom data and all disease symptom data corresponding to the same disease label, including: S34, the first control module substitutes the medical symptom data into the bag-of-words model, and converts all medical terms in the medical symptom data into medical feature vectors; S35. Calculate the cosine value of the angle between the medical consultation feature vector and the standard symptom feature vector.

9. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 1, characterized in that: The data for multiple disease symptoms are fever, cough, headache, abdominal pain, chest tightness, fatigue, loss of appetite, nausea and vomiting, rash and joint pain.

10. The method for monitoring and providing feedback on patient disease information in an AI-based medical guidance system according to claim 1, characterized in that: The multiple treatment recommendations include taking an extended-release ibuprofen tablet twice a day after meals, acupuncture treatment three times a week, and reducing salt intake.

Images