General disease intelligent preliminary screening diagnosis system based on artificial intelligence

Abstract

The application discloses an artificial intelligence-based general disease intelligent preliminary screening diagnosis system, relates to the technical field of medical artificial intelligence, constructs a "data-knowledge-reasoning" closed loop, integrates medical data such as medical records of third-grade class-A hospitals, public health health records and diagnosis and treatment guidelines, extracts key information through natural language processing, converts a clinical path into a structured "disease-feature" correlation library containing symptom intensity coding, and realizes intelligent preliminary screening. The system sets a "medical consultation guidance + result output" module for primary doctors, prompts the collection of chief complaints and generates a disease list with high, medium and low possibility levels and clinical basis; the system sets a "symptom + communication interpretation" module for patients, and outputs diseases, treatment suggestions and typical performance illustrations. The diagnosis suggestion is attached to a traceable reasoning chain, and the system is internally provided with a regional disease library and equipment adaptation rules, and is suitable for the actual situation of primary medical institutions. The system supports dynamic iterative updating of the knowledge base, can be connected to a primary HIS system, realizes two-way calling of health records, and supports continuous diagnosis and treatment.

Description

Technical Field

[0001] This invention relates to the field of medical artificial intelligence technology, specifically to an intelligent initial screening and diagnostic system for general diseases based on artificial intelligence. Background Technology

[0002] As my country's healthcare service system continues to improve, primary healthcare institutions are playing an increasingly prominent role in disease prevention, health management, and the diagnosis and treatment of common diseases. However, primary healthcare faces challenges such as uneven distribution of high-quality medical resources, a shortage of general practitioners, and significant differences in clinical experience. This is especially true when dealing with overlapping symptoms or early manifestations of rare diseases, which can easily lead to missed diagnoses, misdiagnoses, or delayed diagnosis, affecting patients' access to timely and appropriate treatment. At the same time, the public's demand for convenient health consultations continues to grow, but existing self-testing tools for patients often lack medical expertise and fail to provide evidence-based guidance, potentially causing unnecessary panic or delaying medical attention.

[0003] Currently, many hospitals have established electronic medical record (EMR) systems, health record systems, and regional health information platforms, accumulating massive amounts of real-world data such as outpatient medical records, physical examination records, and follow-up information. This multi-source data contains rich clinical characteristics and disease correlation patterns; however, because it is mostly unstructured text and scattered across different systems and institutions, it lacks unified knowledge extraction and computable models, making it difficult to directly use for diagnostic assistance. Traditional clinical decision support systems often rely on fixed rule bases or single guidelines, with long update cycles, and cannot quickly respond to emerging diseases, changes in regionally prevalent diseases, and differences in equipment conditions at the grassroots level, resulting in limited applicability in real-world scenarios.

[0004] In recent years, artificial intelligence, especially natural language processing (NLP) technology, has made progress in medical text analysis, enabling the extraction of key information such as symptoms, signs, and medical history from medical records. However, most research focuses on specific diseases or specialized areas in large hospitals, lacking a unified framework that is applicable to all departments and covers common and frequently occurring diseases. Furthermore, existing intelligent diagnostic systems often exhibit "black box" characteristics in their reasoning process, making it difficult for doctors to trace the basis of their judgments, which is detrimental to clinical trust and continuous optimization. Patient-oriented applications often remain at the level of simple keyword matching or probability output, lacking interpretable explanations and visual aids that incorporate medical logic, thus reducing their practical value.

[0005] In grassroots practice, doctors often struggle to obtain complete chief complaints and medical histories due to limited consultation time and varying patient communication abilities, leading to incomplete information gathering and impacting diagnostic quality. Furthermore, the prevalence of diseases varies significantly across regions. For instance, intestinal infectious diseases and schistosomiasis are more prevalent in southern regions, while cardiovascular diseases and chronic obstructive pulmonary disease are more common in parts of northern China. In contrast, varying levels of medical equipment at the grassroots level may prevent the provision of certain laboratory tests. This necessitates that diagnostic logic possess regional adaptability and feasibility within resource constraints. Existing general-purpose diagnostic models rarely consider these factors, and directly applying them to grassroots settings can easily result in logical biases or impractical outcomes.

[0006] Furthermore, medical knowledge is constantly being updated; treatment guidelines, drug use standards, and disease classification standards are all evolving, and the characteristics of typical cases change with the accumulation of clinical practice. If a system cannot dynamically iterate its knowledge, it will quickly fall behind the needs of practical applications and may even solidify erroneous or outdated reasoning patterns. Most existing systems lack the ability to deeply integrate with medical institution information systems (such as HIS and public health platforms), making it difficult to retrieve and rewrite patient health records and treatment records in real time during the diagnostic process, thus limiting the realization of continuous and personalized services.

[0007] In summary, there is a need for a comprehensive intelligent primary screening and diagnostic system for diseases that can integrate multi-source medical data, transform clinical pathways into quantifiable and structured knowledge, support traceable reasoning processes, take into account the actual conditions and regional characteristics of primary care, and possess continuous learning and system interconnection capabilities. This system would address the shortcomings in current primary care and public health consultation, and improve accessibility and quality of healthcare. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent primary screening and diagnostic system for general diseases based on artificial intelligence. It integrates multi-source medical data to form a structured "disease-feature" association database, providing accurate primary screening and basis for primary care physicians and patients. It combines regional disease databases with equipment adaptation rules to fit the actual situation, and improves the efficiency and continuity of primary care diagnosis through dynamic iteration and HIS.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solution: an intelligent preliminary screening and diagnostic system for general diseases based on artificial intelligence. This system constructs a closed-loop operating architecture of "data acquisition-knowledge organization-logical reasoning", integrates real-world medical data resources from multiple sources, extracts key clinical characterization information from them using text parsing methods, transforms standardized clinical treatment processes into a structured "disease-feature" association knowledge base, and realizes automated preliminary screening and judgment for diseases across the entire medical field based on this knowledge base.

[0010] Furthermore, the multi-source medical data resources specifically include, but are not limited to: structured and unstructured electronic medical records generated by the general practice department or various specialized outpatient clinics in tertiary general hospitals; personal health record data of residents established and maintained in the basic public health service projects organized and implemented by the state; publicly released standardized diagnosis and treatment reference outlines for common diseases; and a set of representative case core features jointly reviewed by a group of clinical experts.

[0011] Furthermore, in the "disease-feature" association knowledge base, each clinical feature associated with a disease is assigned a graded label to indicate the degree of association. For example, for the disease sepsis, the support graded label for a specific combination of clinical features associated with it is higher than the graded label for a combination of features associated with common fever, thus providing a quantifiable reference benchmark for the subsequent logical reasoning process.

[0012] Furthermore, the system is configured to provide services to primary healthcare practitioners, and includes two functional units: "interactive consultation guidance and information collection" and "diagnostic conclusion output and presentation." The "interactive consultation guidance and information collection" unit guides users to gradually collect and confirm the patient's complete main needs and related information through progressive prompts and inquiries. The "diagnostic conclusion output and presentation" unit automatically generates a list containing several candidate diseases and their corresponding "probability levels" based on the collected information, and lists a summary of key clinical information on which the inference is based for each candidate disease.

[0013] Furthermore, in the probability ranking list, each candidate disease is clearly marked with its corresponding "probability level," which is divided into three levels: high, medium, and low. Each level is accompanied by specific clinical evidence, which is a combination of key data and feature descriptions extracted from the collected information. For example, "a hemoglobin concentration of 5 grams per liter was detected, and pale complexion was observed" is used as the basis for inferring moderate anemia.

[0014] Furthermore, the system is also configured to provide information services to non-professional patient users, and has two functional units: "symptom information input and matching analysis" and "simplified translation and communication assistance of results". After the patient user inputs a description of their perceived symptoms through the "symptom information input and matching analysis" unit, the system will analyze and output a set of possible disease indications and corresponding medical treatment directions. Through the "simplified translation and communication assistance of results" unit, typical clinical manifestation diagrams of different possible diseases are attached to help the patient user understand the meaning of the analysis results.

[0015] Furthermore, the diagnostic recommendations output by the system also include a "reasoning path description" that allows for tracing the formation process. The "reasoning path description" clearly lists the core known conditions used to derive the diagnostic recommendation and the matching diagnostic criteria elements, such as "the patient has a history of diabetes, the fasting blood glucose level in this test is 11.2 mmol / L, and the qualitative test result of urine glucose is positive (++). The above conditions together meet the established diagnostic criteria for type 2 diabetes," thereby supporting medical practitioners to retrospectively verify the formation logic of the diagnostic recommendation.

[0016] Furthermore, the system is pre-configured with a "regional common disease characteristic database" and a "primary healthcare equipment capability adaptation rule set". The system can adaptively and dynamically adjust the diagnostic analysis logic based on the spectrum of high-incidence diseases determined by the epidemiological characteristics of different regions in my country, and in combination with the actual configuration level and performance limitations of the existing medical examination and testing equipment in primary healthcare institutions, so as to ensure that the system's diagnostic analysis process is consistent with the actual application scenario.

[0017] Furthermore, the system has a knowledge base and rule set update mechanism based on newly added data; by continuously accessing newly generated clinical observation data and case information, and incorporating confirmed clinical expert experience feedback, the system supplements, corrects, or optimizes the content composition of the "disease-feature" association knowledge base and the applied logical reasoning rules, thereby achieving a gradual improvement in the overall diagnostic accuracy of the system.

[0018] Furthermore, the system is equipped with an interface module for data exchange with the information management system used by primary healthcare institutions. Through this interface module, the system can perform bidirectional querying and synchronous updating of patient personal health record data with the primary healthcare institution information system (HIS), providing data support for the continuous and integrated diagnosis and treatment services for patients in primary healthcare institutions.

[0019] This invention provides an intelligent primary screening and diagnostic system for general diseases based on artificial intelligence, which has the following beneficial effects: 1. This system integrates multi-source data, including medical records from tertiary hospitals, public health archives, and treatment guidelines. It utilizes natural language processing to extract key information and constructs a structured "disease-feature" association database, providing precise intelligent initial screening support for primary care physicians. The "diagnosis guidance + result output" module, designed specifically for primary care physicians, uses prompt-based interactive guidance to collect complete chief complaints and automatically generates a disease list with probability levels and clinical evidence (e.g., "hemoglobin 5g / L + pale complexion" indicates moderate anemia), avoiding missed diagnoses. Simultaneously, it incorporates a "regional disease database" and "equipment adaptation rules," dynamically adjusting diagnostic logic based on regionally prevalent diseases and the available equipment at the primary care level. This allows primary care physicians, with limited resources, to achieve initial screening levels approaching those of higher-level hospitals, effectively alleviating the problem of insufficient diagnostic capacity at the primary care level and shortening patient diagnosis time.

[0020] 2. The system's "Symptom + Communication Interpretation" module not only allows patients to input symptoms and receives possible diseases and treatment suggestions, but also visually demonstrates the differences between various diseases through typical manifestation diagrams (such as a comparison of typical signs of infections and endocrine disorders). This helps patients understand "why this disease should be considered" and "whether emergency medical attention is needed." This visual interpretation breaks the "black box" state of traditional diagnosis, reduces patients' confusion with medical terminology, and avoids duplicate visits or delayed treatment due to information asymmetry. For example, seeing the distribution characteristics of different diseases in the "fever with rash" diagram allows patients to more clearly judge the tendency of their own symptoms, thereby choosing a more appropriate department or time to seek medical attention, improving medical efficiency and cooperation.

[0021] 3. All diagnostic suggestions in the system come with a traceable "chain of reasoning," such as "History of diabetes + fasting blood glucose 11.2 mmol / L + urine glucose++ meets the diagnostic criteria for type 2 diabetes," presenting a complete logical derivation process from medical history to indicators to conclusions. This design allows doctors (especially inexperienced primary care physicians) to quickly backtrack and verify each step of the evidence, avoiding subjective assumptions or omissions of key indicators. Simultaneously, it provides objective evidence for determining liability in medical disputes, enhancing the standardization and credibility of the diagnostic and treatment process. For teaching purposes, young doctors can also learn diagnostic thinking by analyzing the chain of reasoning, accelerating their skill development and contributing to the overall improvement of primary care quality in the long term.

[0022] 4. The system incorporates a targeted "regional disease database" and "equipment adaptation rules," dynamically identifying prevalent diseases in different regions of my country (e.g., dengue fever in the south, COPD in the north) and common equipment limitations at the grassroots level (e.g., alternative examination strategies when CT scanners are unavailable), automatically adjusting diagnostic priorities and verification logic. For example, in regions where only blood and urine tests are available, the system strengthens the correlation weight between blood and urine indicators and diseases, preventing diagnostic failures due to reliance on high-end equipment. This "localized" design allows the system to adapt to real-world grassroots environments without large-scale modifications, solving the problem of general AI tools being "unsuitable for local conditions" and significantly improving the system's implementation rate and application value in county and township-level medical institutions.

[0023] 5. The system possesses dynamic iteration capabilities. By incorporating new clinical data (such as newly added infectious disease cases) and expert feedback, it regularly updates the "disease-feature" association database and inference rules (such as adjusting the support strength of a symptom for a disease), ensuring that the diagnostic model keeps pace with medical advancements. Simultaneously, the system interfaces with primary care HIS (Hospital Information System), supporting bidirectional access and updates to patient health records—doctors can view patients' past physical examinations and medication records, while patients can simultaneously receive updated follow-up recommendations from the system. This closed-loop "data-knowledge-reasoning" system and continuous data linkage not only improve the accuracy of individual diagnoses but also support the long-term management of chronic diseases such as hypertension and diabetes, promoting the transformation of primary healthcare from a "disease-centered" to a "health-centered" approach. Attached Figure Description

[0024] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating the overall "data-knowledge-reasoning" closed-loop construction process of the system of this invention; Figure 2 This is a flowchart of the "consultation guidance + result output" process for primary care physicians in this invention. Figure 3 This is a flowchart of the "symptoms + communication and interpretation" process for patient users in this invention; Figure 4 This is a flowchart of the diagnostic suggestion "chain of reasoning" for this invention; Figure 5 This is a flowchart of the dynamic iteration and data docking process of the system of the present invention. Detailed Implementation

[0026] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses consistent with some aspects of this disclosure as detailed in the appended claims.

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0028] How to use: I. User Process for Primary Care Physicians 1. System Integration and Initialization: After the system is started, it will be integrated with the primary healthcare institution information system (HIS) to enable bidirectional access and updating of patient health records. The regional disease database and equipment adaptation rules will be loaded synchronously. The system will automatically adapt to local high-incidence diseases and existing diagnostic and treatment equipment conditions without manual configuration.

[0029] 2. Consultation guidance and information collection: Select the "Consultation guidance + result output" function module. For patients, follow the system's prompt-based interactive guidance to gradually collect key clinical information such as the patient's chief complaint, symptoms, and medical history. The system automatically extracts core information through natural language processing technology, eliminating the need for manual input of structured data.

[0030] 3. Intelligent Initial Screening and Result Viewing: After information collection is completed, the system calls the "Disease-Feature" association library and, combined with symptom intensity coding and quantitative analysis, automatically generates a preliminary disease screening list. The list marks each disease with a high, medium, or low probability level, along with corresponding clinical evidence, clearly presenting the correlation logic between symptoms and diseases.

[0031] 4. Reasoning Backtracking and Diagnostic Assistance: When viewing the disease list, users can click on the corresponding disease to view the complete reasoning chain, trace back the system's diagnostic basis, and verify the rationality of the diagnosis. If there are any questions about the results, users can supplement the patient's clinical information, and the system will update the initial screening results in real time. After diagnosis, the system will synchronously update the patient's health record to the HIS to support continuous diagnosis and treatment.

[0032] II. User Process for General Patients 1. Functional module selection: After logging into the system, patients can directly access the "Symptoms + Communication Interpretation" functional module without complicated permission verification. The interface design is simple and easy to understand, and it is suitable for non-professional users.

[0033] 2. Symptom Input and Submission: Follow the system prompts to truthfully input your symptoms and related information (such as symptom duration, accompanying symptoms, etc.). After inputting, submit the information. The system will quickly parse the input and match it with disease features in the associated database.

[0034] 3. Result Viewing and Interpretation: The system instantly outputs preliminary screening results, including a list of possible diseases, targeted medical advice, and illustrations of typical symptoms for each disease to help patients intuitively understand disease characteristics and avoid misunderstandings. Patients can choose the corresponding department for treatment based on the advice, improving the efficiency of their medical visit.

[0035] III. General System Operation and Maintenance 1. Iterative Updates: The system has automatic dynamic iteration capabilities, requiring no manual operation from the user. It can continuously access new clinical data and expert feedback, and update the "disease-feature" association library and inference rules in real time, ensuring that diagnostic accuracy gradually improves with data accumulation.

[0036] 2. Adaptation and Adjustment: In response to changes in the regional distribution of diseases, the system automatically updates the regional disease database and adjusts the diagnostic logic; if primary healthcare institutions update their diagnostic and treatment equipment, the system automatically adapts through equipment adaptation rules without the need to reconstruct the diagnostic model.

[0037] 3. Precautions: During use, primary care physicians must ensure the accuracy and completeness of patient information to guarantee the reliability of the reasoning chain; patients must truthfully input their symptom information to avoid inaccurate initial screening results due to information discrepancies. The system only provides initial screening suggestions and cannot replace a doctor's final diagnosis. A definitive diagnosis requires clinical examination and physician judgment.

[0038] Example 1: Implementation of Primary Care Physician Consultation Guidance and Intelligent Initial Screening This embodiment targets primary care physicians, leveraging the system's "diagnosis guidance + result output" module and "disease-feature" association database to achieve standardized diagnosis and accurate initial screening. After seeing a patient, the primary care physician activates the system and completes the connection with the primary healthcare institution's information system (HIS). The system automatically retrieves the patient's past health records and loads the corresponding regional disease database and equipment adaptation rules, adapting to local prevalent diseases and existing treatment conditions, without requiring doctors to manually adjust parameters.

[0039] Doctors select the "Consultation Guidance + Result Output" module. The system guides doctors step-by-step through collecting patient information using a prompt-based interactive approach, including chief symptoms, duration of illness, past medical history, and accompanying symptoms. During this process, the system automatically extracts key clinical information from the unstructured descriptions entered by the doctor using natural language processing technology, eliminating the need for doctors to manually organize structured data and significantly reducing operational complexity. After information collection is complete, the system automatically calls upon its built-in "disease-feature" association database, performs quantitative analysis based on symptom intensity coding, prioritizes matching diseases with high symptom correlation, and generates a preliminary disease screening list with high, medium, and low probability levels.

[0040] Each disease in the list is accompanied by clear clinical evidence, clearly presenting the logical connection between symptoms and the disease. Doctors can click on any disease to view the complete reasoning chain, trace back to the core basis of the system's diagnosis, and verify the rationality of the diagnosis. If doctors have doubts about the initial screening results, they can supplement the patient's clinical information through the system, and the system updates the reasoning logic and initial screening results in real time. After a diagnosis is confirmed, the system automatically synchronizes the treatment information to the HIS, updates the patient's health record, provides support for subsequent continuous treatment, and effectively solves the problems of non-standardized consultations and insufficient accuracy of initial screening by primary care physicians.

[0041] Example 2: Patient Symptom Self-Assessment and Medical Visit Guidance Example This embodiment is designed for general patient users. Through the system's "Symptoms + Communication & Interpretation" module, it facilitates convenient symptom self-assessment and provides scientific guidance for medical treatment, catering to the operational needs of non-professional users. After logging into the system, patients do not require complex permission verification; the interface automatically redirects to the "Symptoms + Communication & Interpretation" module. The module interface features a simplified design, guiding patients through operation with graphic and textual prompts, lowering the barrier to entry.

[0042] Following the system prompts, patients truthfully input their symptoms and related information. The system provides standardized input guidance to avoid information bias caused by ambiguous descriptions. After input and submission, the system quickly parses the patient's input, matches it with disease features in the "disease-feature" association database, and instantly generates preliminary screening results. The results include a list of possible diseases, targeted medical advice, and illustrations of typical manifestations of each disease to help patients understand the core characteristics of the disease in a visual way, avoiding misunderstandings due to a lack of professional knowledge.

[0043] Patients can accurately select the corresponding department for treatment based on the system's suggested visits, reducing the time and energy wasted on aimless visits. Simultaneously, the system uses visual explanations to help patients clearly distinguish the typical symptoms of different diseases, improving their self-health awareness. This embodiment requires no professional medical knowledge from the patient; simple operation is all it takes to obtain scientific initial screening suggestions and treatment guidance, effectively meeting the public's daily health self-check needs.

[0044] Example 3: System Inference Chain Traceability and Diagnostic Verification Example This embodiment focuses on the system's traceable reasoning function. Relying on the "disease-feature" association library and reasoning rules, it enables the diagnostic basis to be traceable and verifiable, ensuring the reliability of the initial screening results. Whether it is a primary care physician or a patient user, after obtaining the system's initial screening results, they can trigger the reasoning chain tracing function, and the system will fully present the entire reasoning logic from information collection to result generation.

[0045] For primary care physicians, tracing the reasoning chain allows them to clearly see the system's symptom extraction process, the matching logic with the "disease-feature" association database, and the quantitative basis for symptom intensity coding. This enables them to accurately determine whether the system's diagnosis aligns with clinical reality, providing a reference for subsequent treatment decisions. If there are links in the reasoning chain that contradict clinical understanding, physicians can supplement key information. The system will then readjust its reasoning rules based on the new information, update the initial screening results, and ensure the rationality of the diagnostic recommendations.

[0046] For patients, visualizing the reasoning chain helps them understand the basis for the system's initial screening results, enhancing their trust in the initial screening recommendations. It also allows them to clearly understand the correlation between their symptoms and the disease, facilitating a more accurate description of their condition to doctors during medical visits. This embodiment, through a traceable reasoning mechanism, breaks down the "black box" problem of artificial intelligence diagnosis, making the diagnostic process transparent and improving the credibility of the system.

[0047] Example 4: System Region Adaptation and Dynamic Iteration Example This embodiment focuses on the system's regional adaptability and dynamic iteration capabilities to achieve deep integration between the system and different grassroots scenarios, while ensuring continuous improvement in diagnostic accuracy. After the system starts, it automatically loads the built-in "regional disease database" and "device adaptation rules," and dynamically adjusts the diagnostic logic based on the disease distribution characteristics of the current usage area, prioritizing matching local high-incidence diseases to ensure that the initial screening results are consistent with the actual regional medical situation.

[0048] When primary healthcare institutions update their diagnostic and treatment equipment, the system automatically identifies the functional scope of the new equipment through equipment adaptation rules and adjusts the adaptation logic of diagnostic indicators without the need to reconstruct the diagnostic model, thus reducing system operation and maintenance costs. Simultaneously, the system possesses automatic dynamic iteration capabilities, requiring no manual user intervention. It continuously integrates new clinical data and expert feedback, updating the content and inference rules of the "disease-feature" association database in real time, continuously optimizing the matching accuracy between symptoms and diseases, and improving diagnostic accuracy.

[0049] For example, when a specific disease shows a high incidence trend in a certain area, the system iteratively updates the data, incorporating the disease's characteristics and related symptoms into a "disease-feature" association database, adjusting the symptom intensity coding logic, and improving the initial screening sensitivity for the disease. This embodiment enables the system to adapt to grassroots scenarios with different regions and equipment conditions, while maintaining advanced diagnostic capabilities through continuous iteration, effectively solving the problems of poor adaptability and difficulty in updating traditional diagnostic systems.

[0050] Example 5: System Integration with HIS and Continuous Medical Treatment Support Example This embodiment focuses on the interface between the system and the primary healthcare institution's information system (HIS), enabling bidirectional transfer of patient health records and supporting the implementation of continuous medical services. After deploying the system, the primary healthcare institution completes the interface configuration with its local HIS system. Once configured, the system can achieve bidirectional retrieval and updating of patient health records, breaking down information silos and realizing the interconnection and interoperability of medical data.

[0051] When doctors use the system to conduct initial screenings for patients, they can directly retrieve patients' past health records and medical records from the HIS system, eliminating the need for patients to provide this information repeatedly. This improves consultation efficiency and provides the system with more comprehensive reference data, enhancing the accuracy of the initial screening. After the initial screening, if the patient is diagnosed, the doctor can synchronize the consultation information, initial screening results, and diagnosis conclusion to the HIS system, automatically updating the patient's health record and ensuring the timeliness and completeness of the information.

[0052] When a patient visits again, the attending physician can access the updated medical records through the HIS system. Combined with the new initial screening results, this allows for a rapid understanding of changes in the patient's condition and treatment history, enabling the development of targeted treatment plans and achieving continuous care. This embodiment, through deep integration of the system and HIS, consolidates medical data resources, reduces redundant work, improves the efficiency of primary healthcare institutions, and provides patients with consistent and accurate medical services, ensuring treatment effectiveness.

[0053] Example 6: Patients use the system to understand the typical manifestations of hand-foot-mouth disease and obtain medical advice. A parent noticed small red rashes on their child's hands and feet, accompanied by drooling and refusal to eat. They logged into the system and accessed the "Symptoms + Communication Interpretation" module, inputting "My child has red rashes on their hands and feet, is drooling, and doesn't want to eat." Based on pre-integrated common disease diagnosis and treatment guidelines and typical case characteristic data, the system analyzed the "Disease-Feature" association database: matching "red rashes on hands and feet + oral discomfort (drooling, refusal to eat)" with the association features of "hand-foot-mouth disease" in the database, outputting "possible disease: hand-foot-mouth disease," and recommending "visit a pediatrician as soon as possible, and carefully observe the child's temperature and mental state." The system also included illustrations of typical hand-foot-mouth disease symptoms—clearly showing the "distribution characteristics of red papules on the palms / soles" and the "location of oral mucosal blisters"—helping the parent visually distinguish between common rashes and the characteristic manifestations of hand-foot-mouth disease, alleviating anxiety, and guiding targeted observation. Because hand-foot-mouth disease has recently been included in the regional disease database for key monitoring in this area, the system has also added to the recommendations, taking into account equipment compatibility rules (there is no rapid pathogen detection equipment at the grassroots level), that "when seeking medical treatment, you can tell the doctor about the duration of symptoms to assist in the initial judgment."

[0054] Example 7: Primary care physicians optimize the initial screening of suspected COVID-19 cases after using the system to dynamically update the knowledge base. A community hospital, acting as a COVID-19 surveillance sentinel, recently integrated new expert-verified case characteristics (including atypical respiratory symptoms following infection with the Omega occulta strain of COVID-19) into its system through dynamic iteration. When a doctor saw a patient with "low-grade fever, decreased sense of smell, and occasional dry cough," the doctor used the "diagnosis guidance" function to ask additional questions such as "Can't smell the aroma of food?" and "Is the cough accompanied by phlegm?" The system collected information that the patient "has been unable to smell coffee for the past two days and has a cough with little phlegm." Based on the updated "disease-feature" association database (which added "decreased sense of smell + dry cough" as supporting features for COVID-19), and combined with the pre-integrated national basic public health service health records (the patient had no history of chronic respiratory diseases), the system initiated inference: generating a disease list, marking "COVID-19 infection (suspected)" as a "medium" probability level, and showing clinical evidence of "low-grade fever + decreased sense of smell + dry cough (common atypical symptoms of COVID-19)." The inference chain traced back to "the patient has low-grade fever (signs of infection) + decreased sense of smell (one of the characteristic manifestations of COVID-19) + dry cough (respiratory symptoms)," which matches the clinical description of a suspected COVID-19 case. The system also uses dynamic iteration capabilities to feed the characteristics of the case back to the knowledge base, assisting in the identification and optimization of similar cases in the future. At the same time, based on the regional disease database (currently COVID-19 is a key monitored disease), the system emphasizes in its recommendations that "antigen testing should be carried out in accordance with the standards and travel history should be disclosed," which is in line with the actual prevention and control scenario.

[0055] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent primary screening and diagnostic system for general diseases based on artificial intelligence, characterized in that: The system constructs a closed loop of "data-knowledge-reasoning", integrates real-world multi-source medical data, extracts key clinical information through natural language processing, and transforms clinical pathways into a structured "disease-feature" association library to achieve intelligent initial screening and diagnosis.

2. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 1, characterized in that, The multi-source medical data includes, but is not limited to, medical records from general practice / specialist outpatient clinics of tertiary hospitals, health records from the National Basic Public Health Service, guidelines for the diagnosis and treatment of common diseases, and typical case characteristics verified by experts.

3. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 1, characterized in that, The "disease-feature" association library contains intensity codes for symptoms, such as a higher support for sepsis than for common fever, providing quantitative evidence for diagnostic reasoning.

4. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 1, characterized in that, The system is designed for primary care physicians and includes a "diagnosis guidance + result output" module. It collects complete chief complaint information through prompt-based interaction and automatically generates a list of diseases including "probability level" and the basis for the diagnosis.

5. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 4, characterized in that, The disease list indicates the "probability level" of each disease as high, medium, or low, and includes clinical evidence. For example, "hemoglobin 5g / L + pale complexion" indicates moderate anemia.

6. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 1, characterized in that, The system is designed for patient users and includes a "symptoms + communication and interpretation" module. After the user inputs their symptoms, the system outputs possible diseases, medical advice, and illustrations of typical symptoms of different diseases to help the user understand them.

7. The intelligent primary screening and diagnostic system for general practice diseases based on artificial intelligence according to claim 1, characterized in that, The system's diagnostic recommendations come with a traceable "chain of reasoning," such as "the patient has a history of diabetes + fasting blood glucose 11.2 mmol / L + urine glucose++" which meets the diagnostic criteria for type 2 diabetes and supports doctors in retrospective verification.

8. The intelligent primary screening and diagnostic system for general diseases based on artificial intelligence according to claim 1, characterized in that, The system has a built-in "regional disease database" and "device adaptation rules" to dynamically adjust the diagnostic logic based on the prevalence of diseases in different regions of my country and the limitations of equipment at the grassroots level, ensuring that it fits the actual scenario.

9. The intelligent primary screening and diagnostic system for general diseases based on artificial intelligence according to claim 1, characterized in that, The system has dynamic iteration capabilities, and can continuously update the "disease-feature" association library and inference rules by incorporating new clinical data and expert feedback, thereby improving diagnostic accuracy.

10. The intelligent primary screening and diagnostic system for general diseases based on artificial intelligence according to claim 1, characterized in that, The system can interface with the primary healthcare information system (HIS) to enable two-way access and updating of patient health records, supporting continuous diagnosis and treatment.

Images