Medical data screening system and screening method based on AI auxiliary inquiry

Abstract

The invention relates to a medical data screening system and method based on AI auxiliary interrogation, the screening system at least comprises a client APP, an AI client and a cloud which are mutually connected, compared with the prior art, the system fully combines the advantages of AI, cloud storage, a user intelligent terminal and an offline diagnosis and treatment end, and realizes data acquisition, classification, fusion and interaction; the upstream of the AI client side can allow a patient to input basic data and image data, the downstream of the AI client side can call a cloud database, pathological data can be accurately screened through feature comparison, and pathological similarity suggestions are provided; in the high-incidence period of infectious diseases, the screening system or the screening method can be used for shunting hospitals, limited medical resources are fully allocated, and cross infection is avoided; by means of big data screening and online and offline diagnosis and treatment combination, the pathology screening accuracy is improved, accurate diagnosis and treatment suggestions can be given, and the medical diagnosis service experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical care information, and in particular to a medical data screening system and screening method based on AI-assisted medical consultation. Background Art

[0002] A common problem faced by domestic hospitals is that during the peak period of infectious diseases (such as influenza A, influenza B, highly pathogenic avian influenza, viral pneumonia, viral hepatitis, dengue fever, etc.), hospitals are usually overcrowded due to limited medical resources, and medical workers are overwhelmed. Some patients have to queue for treatment throughout the entire process, from registration, initial consultation, examination, to follow-up consultation and prescription (or hospitalization), which takes at least half a day and at most two or three days, wasting many patients' precious treatment time. Some patients have even expressed strong dissatisfaction with the hospital and filed complaints and protests.

[0003] To address these challenges, many hospitals have proposed optimizing service processes. For example, the common practice of "check-in priority, late arrivals postponed" prevents patients from exploiting loopholes to cut in line and ensures fairness. Some hospitals have also implemented tiered diagnosis and treatment systems and a first-visit system. Through designated medical and nursing care strategies, patients with mild symptoms are directed to lower-tier hospitals for initial consultation, while those with more complex and severe cases are referred to higher-tier hospitals for follow-up consultations. Another approach is leveraging big data to monitor disease trends and proactively allocate medical resources to prevent concentrated outbreaks of infectious diseases that overwhelm medical institutions. In addition to optimizing diagnosis and treatment processes and protocols, many hospitals are also boldly experimenting with integrating new technologies at the technical level. Over the past two years, with the rise of AI, AI-assisted consultations have begun to be integrated with online diagnosis and treatment, reducing the number of offline consultations and providing a "diversion" for hospitals. AI technology, through natural language processing (NLP), has the potential to rapidly parse patient-uploaded medical records, laboratory reports, and test reports (such as blood counts and pathology slides). Once localized, it can integrate with large databases to automatically annotate abnormal indicators and generate easy-to-understand explanations and recommendations. AI can also enable multimodal data fusion: for example, integrating medical images (such as CT and MRI), text medical records, genetic data, and other sources of information to assist doctors in differential diagnosis of complex diseases. For example, large-scale pathology models can reduce diagnosis time to seconds by analyzing millions of pathology slides. However, AI is not yet integrated with large medical databases, and its functionality is not yet very powerful. Relying solely on AI is likely to result in a high probability of misdiagnosis and misjudgment, which is also irresponsible to patients. Therefore, how to "divert" hospitals during periods of high infectious disease incidence, effectively utilize limited medical resources, improve the accuracy of pathology screening, provide accurate diagnosis and treatment recommendations, and enhance the medical diagnosis service experience, has become a technical challenge that technical personnel in this field urgently need to solve. Summary of the Invention

[0004] To address the medical challenges in the aforementioned background technology, the present invention proposes a medical data screening system and method based on AI-assisted medical consultation. The specific technical solutions are as follows: A medical data screening system based on AI-assisted medical consultation includes at least a client APP, an AI client and a cloud, wherein the client APP and the AI ​​client are both set on the user's smart terminal, the smart terminal has at least a camera module, and the cloud is set at a CDC. The AI ​​client can be locally deployed and called through the client APP. The AI ​​client can call the database on the cloud according to the instructions sent by the client APP, compare the data with the semantic data and image data input by the AI ​​client, match similar cases, perform etiology and lesion analysis, and finally feedback treatment plans and suggestions to the client APP.

[0005] The client APP and AI client are also connected to the offline diagnosis and treatment end for communication, wherein the offline diagnosis and treatment end uses smart terminals such as computers and mobile phones and is controlled by doctors. The offline diagnosis and treatment end is mainly used to determine whether the results of the AI-assisted consultation are accurate when the customer has a first-time diagnosis experience. If it is accurate, the doctor needs to click "Confirm" on the offline diagnosis and treatment end to indicate that the doctor agrees with the results of the AI-assisted consultation, and inform the patient that the treatment plan, medication recommendations or follow-up recommendations given by the AI ​​can be adopted; if the doctor thinks it is correct, the doctor needs to click "Reject" on the offline diagnosis and treatment end to indicate that the doctor does not agree with the results of the AI-assisted consultation, and at the same time feedback the "Reject" information to the patient's client APP, notifying the patient to go to the first-visit hospital for a follow-up visit.

[0006] The AI ​​client can be DeepSeek-R1, Kimi, ChatGPT, Gemini, etc.

[0007] The cloud is used to store case reports, medical knowledge, drug inventory and usage, and patient diagnosis and treatment data reported by each hospital in a case database, a medical knowledge database, a drug library, and patient diagnosis and treatment history data. The case database is stored by case type, forming a branch-like database with classification index codes.

[0008] The medical knowledge database can at least store medical books and journal article data; the medical library stores the types, quantities and storage location data of medicines.

[0009] The patient's diagnosis and treatment history data is coded and stored according to the patient's main body.

[0010] Based on a medical data screening system based on AI-assisted medical interview, a medical data screening method based on AI-assisted medical interview is further proposed. The method includes the following basic steps: Step S1: The user logs in to the client APP, which calls the AI ​​client and prompts the user to enter basic information and briefly describe the symptoms; Step S2: Determine whether step S1 is completed. If so, the client APP calls the camera module, controls the camera to turn on, collects user facial images and tongue coating images, and uploads the collected images frame by frame to the AI ​​client; Step S3: Based on step S2, the client APP asks the customer whether he / she has a first-time doctor experience. If so, the customer is prompted to upload the medical questionnaire, test form or test report, and upload it frame by frame to the AI ​​client through the client APP. If not, no operation is required. Step S4: The AI ​​client analyzes the images uploaded in step S2 and / or step S3, extracts and identifies image features according to the classification rules preset by the client APP, and completes image feature classification based on the identified feature graph; at the same time, it extracts semantic features based on the basic information and disease information input in step S1; Step S5: Based on the classification results of semantic features and image features, the cloud database data is called, and on this basis, case matching is performed to complete the cause and lesion analysis; Step S6: The AI ​​client provides treatment plans, medication / follow-up advice, and the diagnosis and treatment ends.

[0011] In summary, the medical data screening system and screening method based on AI-assisted medical consultation proposed in the present invention have the following beneficial effects compared with the existing technology: This invention fully combines the advantages of AI technology, cloud storage, user intelligent terminals, and offline diagnosis and treatment terminals to achieve data collection, classification, fusion and interaction; The upstream AI client allows users to input basic information and image data, while the downstream AI client can call the cloud database to accurately screen pathology data through feature comparison and provide pathology similarity recommendations. During the peak period of infectious diseases, the use of this screening system or screening method can "divert" hospitals and fully allocate limited medical resources, so that patients who have not yet received a medical consultation or have received a first consultation but not a follow-up consultation can receive their first or follow-up consultation online, thus avoiding cross-infection between these patients and critically ill patients in the hospital; By making full use of limited medical resources, through big data screening, and combining online and offline diagnosis and treatment, the accuracy of pathology screening is improved. This can not only provide accurate diagnosis and treatment recommendations and enhance the user's consultation service experience, but also greatly save the workload of medical staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a block diagram of a medical data screening system based on AI-assisted consultation of the present invention; Figure 2This is a structural diagram of the case matching and etiology and lesion analysis of the present invention. DETAILED DESCRIPTION

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that those skilled in the art will be able to make other improvements or variations without departing from the scope of the present invention, and these improvements or variations are all within the scope of protection of the present invention.

[0014] As attached Figure 1-2 As shown, a medical data screening system based on AI-assisted medical consultation includes at least a client APP, an AI client and a cloud. The client APP and the AI ​​client are both set on the user's (patient's) smart terminal. The smart terminal has at least a camera module. The cloud is set in the CDC. The AI ​​client can be deployed locally and called through the client APP. The AI ​​client can call the cloud database according to the instructions sent by the client APP, compare the data with the semantic data and image data input by the AI ​​client, match similar cases, analyze the cause and lesion, and finally feedback the treatment plan and suggestions to the client APP.

[0015] The above-mentioned client APP and AI client can also communicate with the offline diagnosis and treatment end, where the offline diagnosis and treatment end can be the attending physician's computer, mobile phone or other smart terminal. The offline diagnosis and treatment end is mainly used to determine whether the results of the AI-assisted consultation are accurate when the customer has a first-time diagnosis experience (for example, a consultation form has been issued to the patient at the first visit, the test results have come out and the corresponding test form and test report have been generated). If it is accurate, the doctor needs to click "Confirm" on the offline diagnosis and treatment end to indicate that the doctor agrees with the results of the AI-assisted consultation and inform the patient that the treatment plan, medication recommendations or follow-up recommendations given by the AI ​​can be adopted; if the doctor thinks it is correct, the doctor needs to click "Reject" on the offline diagnosis and treatment end to indicate that the doctor does not agree with the results of the AI-assisted consultation, and at the same time feedback the "Reject" information to the patient's client APP, notifying the patient to go to the first-visit hospital for a follow-up visit.

[0016] The above-mentioned AI client can be DeepSeek-R1, Kimi, ChatGPT, Gemini, etc. As long as the client software can be implanted in a smart terminal or offline diagnosis and treatment terminal, can perform semantic recognition and retrieval, and participate in local parametric model training (such as image feature extraction and classification), this embodiment does not make specific limitations.

[0017] The CDC's cloud-based system stores case databases, medical knowledge databases, pharmaceutical libraries, and patient treatment history data collected from hospitals within its jurisdiction. The case database is categorized by case type. For example, all influenza A cases are classified as Category 2 J10, all viral pneumonia cases as Category 2 A56, all viral hepatitis A cases as Category 2 B15, and all viral hepatitis B cases as Category 2 B16. Category 2 B15 and Category 2 B16 can be further grouped together under Category 1 B, creating a dendritic database with index codes. Based on this, the AI ​​client can search the case database according to the client app's instructions, performing searches and performing similarity associations based on semantic features and classification information extracted from the image feature map.

[0018] In the cloud, the medical knowledge database mainly stores Western medicine books, Chinese medicine theory books (such as Treatise on Febrile Diseases), journal articles published in domestic and foreign medical journals, etc. The medical library mainly involves the types, quantities and storage locations of traditional Chinese medicine and Western medicine, and can be linked with various hospitals and other medical institutions within the jurisdiction. Specifically, the AI ​​client searches the cloud-based medical library to find the types, quantities and storage locations of traditional Chinese medicine and Western medicine in each hospital, follows the map, and rationally allocates medical resources.

[0019] Patient diagnosis and treatment history data is coded and stored according to the patient subject. For example, a patient is a patient subject. No matter which hospital the patient is diagnosed and treated, all previous medical interviews, test sheets, test reports, and electronic medication orders can be stored in the patient's database. Currently, most patients' medical treatment history data are stored by hospital. Patients with the same disease may go to Hospital A for treatment or Hospital B for treatment. All medical treatment history data of patients who go to Hospital A for treatment is stored in Hospital A, and all medical treatment history data of patients who go to Hospital B for treatment is stored in Hospital B. The historical data is scattered. For patients, going to different hospitals for the same disease requires re-issuing medical interviews and tests, which undoubtedly increases their financial burden. In this embodiment, the patient's medical treatment history data is coded and stored in the cloud according to the patient subject. Not only can the AI ​​client easily access it, but patients and doctors can also retrieve the files after entering their identity information and verifying it. This avoids cross-infection caused by all patients rushing into medical institutions during the peak period of infectious diseases, avoids overload of medical institutions, realizes "diversion" of personnel, and improves the medical diagnosis service experience.

[0020] Based on the above content, the present invention further proposes a medical data screening method based on AI-assisted medical consultation, which includes the following basic steps: Step S1: The user logs in to the client APP, which calls the AI ​​client and prompts the user to enter basic information and briefly describe the symptoms; Step S2: Determine whether step S1 is completed. If so, the client APP calls the camera module to control the camera to turn on, collect user facial images, tongue coating images, and images of the affected area, and upload the collected images frame by frame to the AI ​​client; Step S3: Based on step S2, the client APP asks the customer whether he / she has a first-time doctor experience. If so, the customer is prompted to upload the medical questionnaire, test form or test report, and upload it frame by frame to the AI ​​client through the client APP. If not, no operation is required. Step S4: The AI ​​client analyzes the image uploaded in step S2 and / or step S3, extracts and identifies image features according to the classification rules preset by the client APP, and completes image feature classification based on the identified feature map; at the same time, it extracts semantic features based on the basic information and disease information input in step S1; Step S5: Based on the classification results of semantic features and image features, the cloud database data is called, and on this basis, case matching is performed to complete the cause and lesion analysis; Step S6: The AI ​​client provides treatment plans, medication / follow-up advice, and the diagnosis and treatment ends.

[0021] In the above basic steps, step S5 further includes steps S51 to S54: Step S51: Based on the classification results of semantic features and image features, the case database and patient diagnosis and treatment history database in the cloud database are called to screen the medical data and find similar cases for matching; In the above steps, there are generally no less than 5 similar cases, and they are ranked according to the similarity. The similarity is calculated as the percentage obtained by dividing the total number of features by the number of identical features. In addition, similar cases can also provide a list of drugs for treating the disease. Step S52: Based on step S51, search the medical knowledge database according to the case matching result to find the medical explanation of the cause of the disease and the medication; Step S53: Based on step S52, the medicine database is searched in combination with the medicine list and medication, specifically the type, quantity and storage location of the medicine, and medication suggestions and purchase suggestions are given; Step S54: Based on step S53, similar cases, medication recommendations and drug purchase recommendations are output to the case matching unit.

[0022] In the above basic steps, step S6 further includes steps S61-S62: Step S61: If the user has a first-time diagnosis experience, the AI ​​client transmits the treatment plan, medication / follow-up consultation advice information to the offline diagnosis and treatment end, which then reminds the doctor to review the information. Step S62: After the doctor checks, if he thinks that the diagnosis and treatment results given by the AI ​​client are correct, he clicks "Confirm" on the offline diagnosis and treatment terminal, and the diagnosis and treatment ends; if he thinks that the diagnosis and treatment results given by the AI ​​client are wrong, he clicks "Reject" on the offline diagnosis and treatment terminal, and at the same time feedbacks the "Reject" information to the user client APP, notifying the user to go to the first diagnosis hospital for a follow-up visit.

[0023] In summary, the medical data screening system and screening method based on AI-assisted medical consultation of the present invention fully combine the respective advantages of AI technology, cloud storage, user smart terminals, and offline diagnosis and treatment terminals to achieve data collection, classification, fusion and interaction. During the high incidence of infectious diseases, the use of this screening system or screening method can "divert" hospitals and fully allocate limited medical resources, so that patients who have not yet received a medical consultation or have received a first consultation but not a follow-up visit can receive a first consultation or follow-up visit online, avoiding cross-infection between such patients and critically ill patients in the hospital; at the same time, it enables limited medical resources to be "used on the cutting edge", and through big data screening, the combination of online and offline diagnosis and treatment improves the accuracy of pathology screening, can give accurate diagnosis and treatment recommendations, and enhance the medical diagnosis service experience.

[0024] Those skilled in the art should understand that the above embodiments only describe the optimal implementation method of the present invention and are not limited to the above technical solutions. Without departing from the basic concept of the above invention, other technical solutions are formed by any combination of the above technical features or their equivalent features, such as the technical solutions formed by replacing the above technical features with technical features with similar functions invented in the present invention. The scope of protection of the present invention is determined by the technical solutions covered by the claims.

Claims

1. A medical data screening system based on AI-assisted consultation, comprising at least a client app, an AI client, and a cloud, characterized by: The client APP and AI client are both set on the user's smart terminal, and the smart terminal has at least a camera module. The cloud is set at the CDC. The AI ​​client can be locally deployed and called through the client APP. The AI ​​client can call the database on the cloud according to the instructions sent by the client APP, and compare it with the semantic data and image data input by the AI ​​client, match similar cases, perform etiology and lesion analysis, and finally feedback treatment plans and suggestions to the client APP.

2. The medical data screening system based on AI-assisted consultation according to claim 1, characterized in that: The client APP and AI client are also connected to the offline diagnosis and treatment end, which is mainly used to determine whether the results of the AI-assisted diagnosis are accurate when the customer has a first-time diagnosis experience.

3. The medical data screening system based on AI-assisted consultation according to claim 2, characterized in that: The offline diagnosis and treatment end uses smart terminals such as computers and mobile phones and is controlled by doctors.

4. The medical data screening system based on AI-assisted consultation according to claim 1, characterized in that: The cloud is used to store the cases, medical knowledge, medical inventory and usage, and patient treatment reported by each hospital in a case database, a medical knowledge database, a medicine library, and patient treatment history data; among them, the case database is stored by case type, forming a branch-like database with a classification index code.

5. The medical data screening system based on AI-assisted consultation according to claim 4, characterized in that: The medical knowledge database can at least store medical books and journal article data; the medical library stores the types, quantities and storage location data of medicines; and the patient diagnosis and treatment history data is coded and stored according to the patient subject.

6. A medical data screening method based on AI-assisted medical interview, used in the medical data screening system based on AI-assisted medical interview as claimed in claim 1, characterized in that: The method comprises the following basic steps: Step S1: The user logs in to the client APP, which calls the AI ​​client and prompts the user to enter basic information and briefly describe the symptoms; Step S2: Determine whether step S1 is completed. If so, the client APP calls the camera module, controls the camera to turn on, collects user facial images and tongue coating images, and uploads the collected images frame by frame to the AI ​​client; Step S3: Based on step S2, the client APP asks the customer whether he / she has a first-time doctor experience. If so, the customer is prompted to upload the medical questionnaire, test form or test report, and upload it frame by frame to the AI ​​client through the client APP. If not, no operation is required. Step S4: The AI ​​client analyzes the images uploaded in step S2 and / or step S3, extracts and identifies image features according to the classification rules preset by the client APP, and completes image feature classification based on the identified feature graph; at the same time, it extracts semantic features based on the basic information and disease information input in step S1; Step S5: Based on the classification results of semantic features and image features, the cloud database data is called, and on this basis, case matching is performed to complete the cause and lesion analysis; Step S6: The AI ​​client provides treatment plans, medication / follow-up advice, and the diagnosis and treatment ends.

7. The medical data screening method based on AI-assisted interviewing according to claim 6, characterized in that: The step S5 specifically includes the following steps S51 to S54: Step S51: Based on the classification results of semantic features and image features, the case database and patient diagnosis and treatment history database in the cloud database are called to screen the medical data and find similar cases for matching; Step S52: Based on step S51, search the medical knowledge database according to the case matching result to find the medical explanation of the cause of the disease and the medication; Step S53: Based on step S52, the medicine database is searched in combination with the medicine list and medication, specifically the type, quantity and storage location of the medicine, and medication suggestions and purchase suggestions are given; Step S54: Based on step S53, similar cases and medication recommendations are output to the case matching unit. and drug purchasing advice.

8. The medical data screening method based on AI-assisted interviewing according to claim 6, characterized in that: The step S5 specifically includes the following steps S61-S62: Step S61: If the user has a first-time diagnosis experience, the AI ​​client transmits the treatment plan, medication / follow-up consultation advice information to the offline diagnosis and treatment end, which then reminds the doctor to review the information. Step S62: After the doctor checks, if he thinks the diagnosis and treatment results given by the AI ​​client are correct, he clicks "Confirm" on the offline diagnosis and treatment terminal, and the diagnosis and treatment ends; if he thinks the diagnosis and treatment results given by the AI ​​client are wrong, he clicks "Reject" on the offline diagnosis and treatment terminal, and at the same time feedbacks the "Reject" information to the user's client APP, notifying the user to go to the first diagnosis hospital for a follow-up visit.

Images