Online consultation service-based accompanying diagnosis system

Abstract

The invention discloses an accompanying diagnosis system based on an online consultation service, relates to the technical field of medical data processing, and solves the problems that a traditional accompanying diagnosis system is inaccurate in demand classification and low in triage efficiency and an online consultation demand is difficult to convert into an effective outpatient service. The patient demand acquisition module extracts a feature vector Fpd of diagnosis related data, and the demand classification module determines demand splitting and matching adjustment by calculating a composite demand association degree Fgx; the doctor-patient matching module optimizes a doctor recommendation sequence by calculating a dynamic triage adjustment coefficient Fdz; the demand conversion module calls a hospital business system API to query medical resources, counts a successful scheduling frequency Rac, a total scheduling attempt frequency Rat and a resource availability rate Rar to calculate a resource scheduling success rate Rds, and adjusts demand distribution according to the Rds; the demand conversion rate Ctr is calculated, and the matching strategy, the reservation mechanism and the patient guidance are optimized according to the Ctr, so that the success rate of conversion from the online demand to the offline service is improved, and the efficiency of the accompanying diagnosis system and the patient diagnosis experience are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a companion diagnosis system based on online consulting services. Background Art

[0002] The development of the companion diagnosis system stems from the reform needs of medical informatization and hierarchical diagnosis and treatment, aiming to optimize the efficiency of doctor-patient matching through technical means. In the early days, it mainly relied on manual triage, which had problems such as strong subjectivity and low efficiency. In the early 21st century, with the popularization of electronic medical record systems, primary companion diagnosis systems based on rule engines emerged, which can achieve simple department recommendations. In recent years, breakthroughs in artificial intelligence technology have enabled the system to handle more complex medical needs matching, but existing technologies still have significant defects; With the development of medical information technology, in outpatient clinics, the various medical needs of patients and the medical services provided by hospitals cannot be accurately matched, resulting in a large number of patients piling up in outpatient clinics, visiting the wrong departments, finding the wrong doctors, and following the wrong procedures. This not only wastes medical resources, but also reduces the patient's medical experience and even affects their condition.

[0003] In addition, the traditional triage method relies on department division, which cannot adapt to the differences in department naming, overlapping department business, and the complexity of physician subspecialties in different hospitals, resulting in inaccurate triage and low efficiency in resource allocation. Finally, traditional systems find it difficult to effectively convert online consultation needs into offline business needs when processing patient needs. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a companion diagnosis system based on online consultation services, which solves the technical shortcomings mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a companion diagnosis system based on online consultation services, including a patient demand collection module, a demand classification module, a rule filtering module, a doctor feature vector library module, a doctor-patient matching module and a correction and adjustment module; The patient demand collection module collects patient consultation data through the mobile terminal and uses natural language processing to extract the patient demand feature vector Fpd; The composite demand identification and decomposition module receives the patient demand feature vector Fpd and identifies whether it is a composite demand. The patient demand feature vector Fpd includes information inquiry, transactional demand, health consultation and medical consultation; When the demand feature vector Fpd contains at least two demands, it is identified as a composite demand, and the large language model LLM technology is used to decompose the patient demand feature vector Fpd into multiple independent single feature demands; A rule filtering module, which is used to build predefined rules, including pediatric age restrictions, gender department restrictions, and emergency priority, to filter unreasonable matches; A doctor feature vector library module, which is used to build doctor feature vectors Fys of candidate doctor specialties and service feature indicators, where the doctor feature vector Fys is generated by topic modeling of historical valid consultation data; A doctor-patient matching module, which is used to calculate the matching degree Rm between each independent feature requirement in the patient demand feature vector Fpd and the doctor feature vector Fys, and sort and recommend; at the same time, the consultation request is delivered to the grab pool, and doctors who meet the matching conditions can take the order actively; A correction and adjustment module, which is used to evaluate the resource scheduling success rate Rds, adjust the demand conversion process, and provide an interface for the doctor to maintain the feature corpus.

[0006] Preferably, the demand classification module includes a demand recognition unit and a demand correlation analysis unit; After receiving the consultation-related data, the demand recognition unit extracts the key information in the text by performing text parsing, including word segmentation, part-of-speech tagging, syntactic analysis, and entity recognition; then, based on a deep learning neural network, a semantic classification model is constructed and trained, and the consultation-related data is classified by the semantic classification model, and the semantic similarity Fy between the input text and the demand category is calculated. Finally, information queries, transactional demands, health consultations, and medical consultations are identified, and different demand categories are split based on semantic similarity and association rule mining. By analyzing historical patient consultation data, the co-occurrence relationship of different demand categories is statistically analyzed, the demand co-occurrence probability Fg is obtained, and the corresponding demand category labels are constructed; finally, by tracking the context information of the patient's consultation text and analyzing the content changes of the patient's previous and subsequent inputs, the context relevance Fs is obtained.

[0007] Preferably, the demand correlation analysis unit is used to analyze the logical relationship between each demand category, and perform demand adjustment and matching adjustment; Through semantic similarity analysis, co-occurrence relationship mining, and context correlation analysis, the consultation data of the patient's online consultation is obtained, and the consultation data is semantically analyzed and classified to obtain a consultation classification result. The consultation classification result is matched and analyzed with the preprocessed historical patient consultation data to calculate the semantic similarity Fy, the demand co-occurrence probability Fg, and the context relevance Fs respectively. After dimensionless processing and fitting, the correlation degree Fgx of the composite demand is calculated. The specific calculation formula is as follows: ; Based on the historical patient demand classification data, the average correlation degree of different demand categories is extracted through statistical analysis and decision tree machine learning models, and the correlation threshold F1 is preset; the correlation threshold F1 is compared and evaluated with the correlation degree Fgx, and demand adjustment and matching adjustment are carried out; the specific evaluation content is as follows: When the correlation degree Fgx ≥ the correlation threshold F1, it is determined that the composite demand in the current consultation-related data does not need to be split, and the integrity is retained, and a comprehensive doctor or department including multiple categories is recommended during the matching process; When the correlation degree Fgx < the correlation threshold F1, it is determined that the composite demand in the current consultation-related data needs to be split, and each demand is respectively matched to the corresponding doctor or department, and at the same time, the personalized recommendation adjustment of the patient's demand is carried out.

[0008] Preferably, the rule filtering module intelligently screens the patient's needs based on a preset medical business rule library; first, an age recognition mechanism is established to analyze the patient's age information. When it is detected that the patient's age is less than 14 years old and the disease type does not belong to the trauma category, the pediatric allocation rule is automatically triggered, and the current patient's needs are directed to the pediatric diagnosis and treatment queue; Secondly, a gender and department matching mechanism is established and passed, and the patient's gender information and the department's reception scope are cross-validated. When a male patient is identified, the matching path of obstetricians and gynecologists is automatically blocked; Finally, an emergency grading mechanism is established to evaluate the priority of the patient's condition. When it is determined that the patient is in critical and severe condition, the priority matching mechanism is automatically activated, and the patient's needs are pushed to the emergency treatment channel of the corresponding specialty; In addition, the age recognition mechanism, gender and department matching mechanism, and emergency grading mechanism in the rule filtering module achieve composite condition judgment through logical AND operations.

[0009] Preferably, during the operation of the doctor feature vector library module, a dual-path matching mechanism is adopted. On the one hand, based on the LDA algorithm, topic modeling is carried out on the desensitized historical consultation data, the topic feature vectors of doctors are constructed, and service class indicators are counted; on the other hand, the patient's consultation corpus is stored in a featureized manner, and the most relevant N historical corpora are retrieved and matched by similarity during new consultations, and doctors are recommended based on the distribution of the doctors to which the corpora belong using a voting mechanism; At the same time, a doctor maintenance interface is set up for doctors to regularly review and correct feature data. Finally, recommendations are generated by weighted fusion of the matching results of the topic distribution path and the corpus similarity path, where the weight of the topic path increases with the accumulation of corpora, while the weight of the similarity path remains at a relatively high proportion in the initial stage of new topics to ensure matching accuracy; Meanwhile, service indicators such as doctors' patient satisfaction, consultation response rate, and response timeliness rate are statistically calculated in real time; in addition, a doctor maintenance interface is established to allow doctors to regularly review and correct the topic classification in the feature corpus, thereby adjusting the professionalism and timeliness of the feature vectors, and finally outputting comprehensive doctor feature data including multi-dimensional professional feature vectors and quantified service indicators.

[0010] Preferably, the doctor-patient matching module includes a matching calculation unit, a triage optimization unit, and a privacy protection unit; The matching calculation unit is used to calculate the matching degree Rm of each patient's need based on the patient need feature vector Fpd and the doctor feature vector Fys; first, the matching degree Rm of the patient's need is defined as the specific calculation formula as follows: ; In the formula, represents the dot product of the patient need feature vector and the doctor feature vector; represents the Euclidean norm of the patient need feature vector, and the patient need feature vector Fpd consists of multiple need-related features; Finally, the heap sort algorithm is used to sort the doctor recommendation order from high to low according to the value of the matching degree Rm; after the sorting is completed, the system recommends the doctor with the highest matching degree according to the sorting result; when the patient does not specify a doctor, or the doctor who meets the matching degree Rm is unable to receive the patient, the consultation request is delivered to the grab pool, and the doctor who meets the matching conditions takes the initiative to receive the order, and the detailed information of the recommended doctor, including professional background, historical reception data, and desensitized similar consultation cases, is displayed to the patient for the patient to choose independently; Meanwhile, combined with the hospital scheduling system and the statistics of doctors' activity levels at different times, the possibility of doctors providing services currently is evaluated, and a matching degree adjustment factor Fa is set to affect the final sorting result, and its calculation formula is as follows: ; Among them, is the adjusted matching degree; Fa is calculated by statistically analyzing the online rate, reception situation, and response time of doctors at different times.

[0011] Preferably, the triage optimization unit is used to calculate, through the electronic medical record system, the cross-department matching index Ksp by statistically analyzing the number of departments involved in the patient and the coverage range of the doctor's diagnosis and treatment departments, and through the hospital scheduling system, the remaining available reception slots of the doctor currently are obtained in real time and analyzed to obtain the doctor resource utilization rate Dzy with the maximum reception capacity; Then, the demand matching degree Rm, the cross-department matching index Ksp, and the doctor resource utilization rate Dzy are extracted and dimensionless processed, and the dynamic triage adjustment coefficient Fdz is calculated through the following formula: ; The dynamic triage adjustment threshold F2 is read and preset based on historical matching data analysis, patient satisfaction feedback, and doctor reception load balancing strategy, and compared and evaluated with the dynamic triage adjustment coefficient Fdz. The specific evaluation contents are as follows: If the dynamic triage adjustment coefficient Fdz ≥ the dynamic triage adjustment threshold F2, it means that the current matching scheme meets the triage requirements, and the patient is matched to the current doctor at this time; If the dynamic triage adjustment coefficient Fdz is less than the dynamic triage adjustment threshold F2, it means that the current matching scheme does not meet the triage requirements. In this case, the dynamic triage adjustment includes: Adjust the cross-department matching tolerance Ksp to allow more doctors who meet some requirements to be included in the candidate list; Combined with the doctor resource utilization rate Dzy, patients are guided to be downward compatible according to the doctor recommendation order and matched with doctors with spare consultation capacity; Finally, based on the doctor's real-time workload, including the number of online consultations, and the conflict status of surgery or offline outpatient clinics, multi-dimensional prompt information is generated, including: the response rate of the current period, the number of patients seen during the same period, and the estimated waiting time; the system sets the minimum threshold of the professional matching degree Rm as a screening prerequisite, allowing patients to sort and screen candidate doctors based on multiple dimensions such as response probability and waiting time, and at the same time supports doctors to set personalized consultation rules and reception capacity limits through the maintenance interface.

[0012] Preferably, the privacy protection unit is used to desensitize the patient's medical treatment-related data, including de-identification, pseudo-anonymization and encrypted data storage; first, by adopting a hierarchical data desensitization strategy, the desensitization method of data at different levels is determined, and the patient's identity information including the patient's name, ID number and contact information is stored using a hash function or irreversible encryption; for the patient's medical records and diagnostic information, partial character masking is used, including character replacement of key fields in the medical record number and medical record, while maintaining the integrity of the data format; for text information involving the patient's privacy, including the chief complaint description and past medical history.

[0013] Preferably, the demand conversion module includes a resource scheduling unit and a demand assessment and optimization unit; The resource scheduling unit first receives the patient's online consultation needs, analyzes the demand type, and identifies whether it involves face-to-face consultation, examination, or hospitalization arrangements; then, by calling the hospital business system API interface, it queries and analyzes the current available status of medical resources, including doctor scheduling, examination equipment idle time, and bed occupancy; after obtaining the medical resource status, the system executes resource scheduling attempts. During the scheduling execution process, the number of successful scheduling attempts Rac, the total number of scheduling attempts Rat, and the resource availability rate Rar are recorded and collected. After dimensionless processing, the resource scheduling success rate Rds is calculated. The specific calculation formula is as follows: ; Based on the historical scheduling data of medical resources, calculate the optimal scheduling success rate threshold under different medical scenarios through statistical analysis and regression analysis machine learning models, preset the resource scheduling success rate threshold R, and compare and evaluate it with the resource scheduling success rate Rds. The specific evaluation content is as follows: When the resource scheduling success rate Rds ≥ the resource scheduling success rate threshold R, it indicates that the resource scheduling is successful, the current medical resources meet the requirements, and the system executes business processes such as face-to-face consultation, examination, and hospitalization arrangement according to the established demand allocation strategy; at the same time, it enters the calculation link of the demand conversion rate Ctr to evaluate the proportion of online demands successfully converted into offline consultations. When the resource scheduling success rate Rds < the resource scheduling success rate threshold R, it indicates that the resource scheduling is unsuccessful, and the current medical resources do not meet the requirements. The reasons include that due to overbooking of doctors, tight examination schedules, or insufficient hospital beds, some demands cannot be scheduled; at this time, the system needs to adjust the demand allocation strategy, including: re-matching available resources, guiding patients to other medical institutions or departments; And only after the adjustment, the resource scheduling success rate Rds is re-evaluated until it is greater than or equal to the resource scheduling success rate threshold R, then the system enters the calculation and evaluation link of the demand conversion rate Ctr, otherwise, the scheduling strategy continues to be adjusted.

[0014] Preferably, the demand evaluation and optimization unit is responsible for calculating and evaluating the demand conversion rate Ctr; based on the actual medical records of patients in the hospital business system, including the proportion statistics of the number of reserved face-to-face consultations, the number of completed examinations, and the number of hospitalization arrangements in the online consultation demands, obtain the face-to-face consultation proportion Cmz, the examination proportion Cjc, and the hospitalization arrangement proportion Czh; After extracting the face-to-face consultation proportion Cmz, the examination proportion Cjc, and the hospitalization arrangement proportion Czh and performing dimensionless processing, the demand conversion rate Ctr is calculated through the following formula:

[0015] Based on the historical conversion data, calculate the minimum acceptable conversion rate under different medical scenarios through statistical analysis and clustering analysis machine learning models, preset the demand conversion rate threshold C, and compare and evaluate it with the demand conversion rate Ctr. The specific evaluation content is as follows: When the demand conversion rate Ctr ≥ the demand conversion rate threshold C, it indicates that the demand conversion rate is qualified, the conversion process is normal, and the proportion of online consultations successfully converted into offline services reaches or exceeds the preset threshold, indicating that the system's companion diagnosis process, resource scheduling, and patient demand matching are all at a normal level; When the demand conversion rate Ctr < the demand conversion rate threshold C, it indicates that the demand conversion rate is unqualified and the conversion process is abnormal. The reasons for the unqualified demand conversion rate include factors such as the failure to complete the appointment smoothly, the patient's abandonment of the face-to-face consultation, the conflict in the examination scheduling, or the shortage of hospital beds. At this time, the system needs to adjust the demand conversion process, including: adjusting the matching strategy, improving the appointment mechanism, and adjusting the patient guidance. After optimization and adjustment, re-evaluate the demand conversion rate Ctr until the demand conversion rate is qualified.

[0016] The present invention provides a companion diagnosis system based on online consultation services. It has the following beneficial effects: (1) This companion diagnosis system based on online consultation services effectively solves the problem that traditional companion diagnosis systems cannot accurately distinguish the types of patient needs. By obtaining consultation-related data through the patient need acquisition module, and extracting the patient need feature vector Fpd based on natural language processing technology, and then classifying and processing by the need classification module to identify information query, transactional needs, health consultation, and medical consultation, and calculating the correlation degree Fgx of the composite needs in the consultation-related data, and using the semantic similarity Fy, the need co-occurrence probability Fg, and the context relevance Fs for fitting calculation to decide whether to split the composite needs and make personalized recommendation adjustments. By constructing the doctor feature vector Fys through the doctor feature vector library module, and calculating the patient need matching degree Rm in the doctor-patient matching module, using the matching calculation unit to perform the dot product operation and Euclidean norm calculation of Fpd and Fys for matching sorting, and combining the triage optimization unit to calculate the cross-department matching parameter Ksp and the doctor resource utilization rate Dzy, and finally optimizing the doctor recommendation order by calculating the dynamic triage adjustment coefficient Fdz to improve the accuracy of doctor-patient matching and the rationality of need matching. (2) This companion diagnosis system based on online consultation services solves the problems of inaccurate triage and low resource allocation efficiency caused by traditional triage methods. Through the triage optimization unit of the doctor-patient matching module, combined with the electronic medical record system and the hospital scheduling system, obtain the number of departments involved by the patient and the coverage range of the doctor's diagnosis and treatment departments, calculate the cross-department matching index Ksp, and calculate the doctor resource utilization rate Dzy based on the remaining available appointment slots and the maximum appointment capacity of the doctor. Then extract Rm, Ksp, and Dzy for dimensionless processing, and perform triage optimization and adjustment by calculating the dynamic triage adjustment coefficient Fdz. When Fdz ≥ F2, the system maintains the current matching plan to ensure that the patient is matched to the current doctor. When Fdz < F2, the system dynamically adjusts the triage strategy, including adjusting Ksp to allow more doctors who meet part of the needs to be included in the candidate list, guiding the patient to match with doctors with available appointment capabilities in combination with Dzy, and recommending doctors with multi-disciplinary diagnosis and treatment capabilities to optimize resource allocation and improve the accuracy of triage and the doctor's appointment efficiency. (3)The companion diagnosis system based on online consultation services solves the problem that traditional systems are difficult to effectively convert online consultation needs into offline business needs; through the resource scheduling unit of the demand conversion module, it analyzes the online consultation needs of patients, calls the API interface of the hospital business system to query the doctor's schedule, the idle time of inspection equipment, and the occupancy of hospital beds, attempts resource scheduling after obtaining the status of medical resources, counts the number of successful scheduling times Rac, the total number of scheduling attempts Rat, and the resource availability rate Rar, calculates the resource scheduling success rate Rds and compares it with the resource scheduling success rate threshold R; when Rds, it executes business processes such as face-to-face consultation, examination, and hospitalization arrangement, and calculates the demand conversion rate Ctr, where the demand conversion rate Ctr is calculated from the face-to-face consultation ratio Cmz, the examination ratio Cjc, and the hospitalization arrangement ratio Czh; and according to the evaluation content of the demand conversion rate Ctr, it indicates that the conversion process is normal, otherwise the system adjusts the matching strategy, optimizes the reservation mechanism, and adjusts patient guidance until the demand conversion rate Ctr is qualified, thereby improving the conversion success rate of online consultation needs to offline business, and enhancing the overall efficiency of the companion diagnosis system and the patient's medical experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic framework structure diagram of a companion diagnosis system based on online consultation services according to the present invention; Figure 2 is a schematic logical structure diagram of the demand matching degree Rm in a companion diagnosis system based on online consultation services according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] Embodiment 1 Please refer to Figure 1 , the present invention provides a system including a patient demand collection module, a demand classification module, a rule filtering module, a doctor feature vector library module, a doctor-patient matching module, and a correction and adjustment module; The patient demand collection module collects patient medical data through the mobile terminal and extracts the patient demand feature vector Fpd using natural language processing; The composite demand recognition and decomposition module receives the patient demand feature vector Fpd and identifies whether it is a composite demand. The patient demand feature vector Fpd includes information query, transactional demand, health consultation, and medical consultation; When the demand feature vector Fpd contains at least two types of demands, it is identified as a composite demand, and the large language model (LLM) technology is used to decompose the patient demand feature vector Fpd into multiple independent single feature demands; The rule filtering module is used to build in predefined rules, including pediatric age limits, gender department limits, and emergency priority, to filter out unreasonable matches; The doctor feature vector library module is used to construct doctor feature vectors Fys and service - type feature indicators for the candidate doctor specialties, where the doctor feature vector Fys is generated through topic modeling of historical valid consultation data; The doctor - patient matching module is used to calculate the matching degree Rm between each independent feature demand in the patient demand feature vector Fpd and the doctor feature vector Fys, and sort and recommend; at the same time, the consultation request is delivered to the grab - order pool, and doctors who meet the matching conditions can actively take the order; The correction and adjustment module is used to evaluate the resource scheduling success rate Rds, adjust the demand conversion process, and provide an interface for the doctor to maintain the feature corpus.

[0020] In this embodiment, after the patient demand acquisition module obtains the data related to seeking medical treatment, it uses a combination of rule matching and machine learning to verify the data, including field integrity check, format standardization verification, outlier detection, and time - series consistency analysis, and performs data cleaning, format standardization, and invalid data filtering; based on natural language processing technology, it analyzes the medical treatment data, extracts patient demand information, and after text pre - processing, semantic feature extraction, feature vector construction, and dimensionality reduction, generates the patient demand feature vector Fpd; effectively identify information queries, transactional demands, health consultations, and medical consultations, and reasonably split or integrate composite demands to improve the accuracy of demand classification; The doctor feature vector library module constructs and dynamically updates the doctor feature vector Fys based on doctor business - feature - related data, including consultation records, areas of expertise, and patient evaluations, to ensure the timeliness and matching accuracy of doctor information; among them, the collection time of historical valid consultation data is at least three months before the data collection date and at most six months before the data collection date, depending on the actual situation.

[0021] The doctor-patient matching module calculates the patient demand matching degree Rm based on the patient demand feature vector Fpd and the doctor feature vector Fys, optimizes the doctor recommendation order, and combines the cross-department matching parameter Ksp and the doctor resource utilization rate Dzy to calculate the dynamic triage adjustment coefficient Fdz, optimize the triage strategy, and improve the accuracy of doctor-patient matching; the demand conversion module calculates the resource scheduling success rate Rds, queries the doctor's schedule, the idle time of inspection equipment, and the occupancy of hospital beds through the hospital business system API interface to ensure the reasonable allocation of medical resources, and calculates the demand conversion rate Ctr, optimizes the matching strategy, adjusts the appointment mechanism and patient guidance, improves the success rate of converting online consultation needs into offline services, optimizes the accompanying diagnosis process, and improves the utilization efficiency of medical resources and the patient's medical experience.

[0022] Embodiment 2 The demand classification module includes a demand recognition unit and a demand correlation analysis unit; After receiving the data related to seeking medical treatment, the demand recognition unit extracts the key information in the text through text parsing, including word segmentation, part-of-speech tagging, syntactic analysis, and entity recognition; then, based on the deep learning neural network, a semantic classification model is constructed and trained, and the data related to seeking medical treatment is classified through the semantic classification model, and the semantic similarity Fy between the input text and the demand category is calculated. Finally, information query, transactional needs, health consultation, and medical consultation are identified, and different demand categories are split based on semantic similarity and association rule mining. By analyzing the historical patient consultation data, the co-occurrence relationship of different types of demands is statistically analyzed to obtain the demand co-occurrence probability Fg, and the corresponding demand category labels are constructed; finally, by tracking the context information of the patient's text for seeking medical treatment and analyzing the content changes of the patient's previous and subsequent inputs, the context relevance Fs is obtained.

[0023] The demand correlation analysis unit is used to analyze the logical relationship between various demand categories and perform demand adjustment and matching adjustment; Through semantic similarity analysis, co-occurrence relationship mining, and context correlation analysis, the consultation data of the patient's online consultation is obtained, and the consultation data is semantically analyzed and classified to obtain the consultation classification result. The consultation classification result is matched and analyzed with the preprocessed historical patient consultation data to calculate the semantic similarity Fy, the demand co-occurrence probability Fg, and the context relevance Fs respectively. After dimensionless processing and fitting, the correlation degree Fgx of the composite demand is calculated. The specific calculation formula is as follows: ; Based on the historical patient demand classification data, the average correlation degree of different demand categories is extracted through statistical analysis and the decision tree machine learning model, and the correlation degree threshold F1 is preset; the correlation degree threshold F1 is compared and evaluated with the correlation degree Fgx to perform demand adjustment and matching adjustment; the specific evaluation content is as follows: When the correlation degree Fgx ≥ the correlation degree threshold F1, it is determined that the composite requirements in the current consultation-related data do not need to be split, and the integrity is retained. During the matching process, comprehensive doctors or departments including multiple categories are recommended; When the correlation degree Fgx < the correlation degree threshold F1, it is determined that the composite requirements in the current consultation-related data need to be split. Each requirement is respectively matched to the corresponding doctor or department, and at the same time, personalized recommendation adjustment of the patient's requirements is carried out.

[0024] The rule filtering module conducts intelligent screening of the patient's requirements based on a preset medical business rule library; first, an age recognition mechanism is established to analyze the patient's age information. When it is detected that the patient's age is less than 14 years old and the disease type does not belong to the trauma category, the pediatric allocation rule is automatically triggered, and the current patient's requirements are directed to the pediatric diagnosis and treatment queue; Secondly, a gender and department matching mechanism is established and passed. The patient's gender information and the department's reception scope are cross-validated. When it is identified as a male patient, the matching path of obstetricians and gynecologists is automatically blocked; Finally, an emergency grading mechanism is established to evaluate the priority of the patient's condition. When it is determined that the patient is an emergency critical patient, the priority matching mechanism is automatically activated, and the patient's requirements are pushed to the emergency treatment channel of the corresponding specialty; In addition, the age recognition mechanism, gender and department matching mechanism, and emergency grading mechanism in the rule filtering module achieve composite condition judgment through logical AND operation.

[0025] During the operation of the doctor feature vector library module, a dual-path matching mechanism is adopted. On the one hand, based on the LDA algorithm, topic modeling is carried out on the desensitized historical consultation data to construct the doctor's topic feature vector and count service indicators; on the other hand, the patient's consultation corpus is characterized and stored. When there is a new consultation, the most relevant N historical corpora are retrieved and matched through similarity, and doctors are recommended based on the distribution of the doctors to which the corpora belong using a voting mechanism; At the same time, a doctor maintenance interface is set up for doctors to regularly review and correct feature data. Finally, recommendations are generated by weighted fusion of the matching results of the topic distribution path and the corpus similarity path. Among them, the weight of the topic path increases with the accumulation of corpora, while the weight of the similarity path remains at a relatively high proportion in the initial stage of new topics to ensure the matching accuracy; At the same time, service indicators including the patient satisfaction, consultation response rate, and response timeliness rate of doctors are statistically calculated in real time; in addition, a doctor maintenance interface is set up to allow doctors to regularly review and correct the topic classification in the feature corpus, so as to adjust the professionalism and timeliness of the feature vector. Finally, comprehensive doctor feature data including multi-dimensional professional feature vectors and quantified service indicators is output.

[0026] In this embodiment, the requirement classification module includes a requirement recognition unit and a requirement correlation analysis unit. The requirement recognition unit constructs a semantic classification model based on deep learning, and calculates the semantic similarity Fy (the semantic matching degree between the patient input text and the preset requirement category label) to identify information queries, transactional requirements, health consultations, and medical consultations. At the same time, the co-occurrence probability Fg of requirements (the conditional probability of different requirement categories appearing simultaneously in historical data) is obtained through association rule mining, and the context relevance Fs (the logical coherence of requirement changes in the patient's consecutive conversations) is calculated by tracking the context information of the patient's consultation text. Finally, the composite requirement correlation degree Fgx is calculated in the requirement correlation analysis unit, and evaluated according to the correlation threshold F1 to determine the splitting or merging of composite requirements, optimizing the matching accuracy; among them, the association rule specifically adopts the FP-Growth algorithm, which specifically analyzes the historical patient consultation data to mine the frequent co-occurrence patterns and conditional probability relationships between different requirement categories. For example, when the patient's main complaint is "stomachache", the requirements often include "medical consultation (department of gastroenterology)" and "transactional requirement (gastroscopy appointment)" at the same time. The doctor feature vector library module collects data related to the business characteristics of doctors through the hospital information management system, electronic medical record system, and online consultation platform, and performs data cleaning, format conversion, and outlier removal. Subsequently, natural language processing technology and multi-dimensional feature extraction methods are used to analyze the doctor's diagnosis and treatment texts, patient evaluation contents, and diagnosis disease distributions, extract the doctor's professional characteristics, service capabilities, and patient satisfaction, and construct the doctor feature vector Fys. Among them, the doctor's reception records are used to extract the doctor's historical consultation types and treatment tendencies, and the areas of expertise are classified and calculated through academic papers, department affiliations, and doctors' self-declared information. The patient evaluation data is quantified based on sentiment analysis and score normalization methods. Finally, a dynamic update mechanism is used to regularly incrementally adjust Fys at fixed times and perform real-time optimization in combination with the latest data to improve the accuracy and timeliness of doctor matching; this system can accurately identify patient requirements, intelligently classify and optimize requirement matching, improve the rationality and accuracy of doctor-patient matching, and at the same time optimize medical resource scheduling, enabling patient requirements to be effectively transformed into offline services, improving hospital resource utilization and patient visit experience, and enhancing the intelligence and accuracy level of the companion diagnosis system.

[0027] Embodiment 3 The doctor-patient matching module includes a matching calculation unit, a triage optimization unit, and a privacy protection unit; The matching calculation unit is used to calculate the matching degree Rm of each patient requirement based on the patient requirement feature vector Fpd and the doctor feature vector Fys; first, the matching degree Rm of the patient requirement is defined as the specific calculation formula as follows: ; In the formula, Denotes the dot product of the patient demand feature vector and the doctor feature vector; Denotes the Euclidean norm of the patient demand feature vector, and the patient demand feature vector Fpd consists of multiple demand-related features; Finally, the heap sort algorithm is adopted to sort the doctor recommendation order in descending order according to the value of the matching degree Rm; after the sorting is completed, the system recommends the doctor with the highest matching degree according to the sorting result; when the patient does not specify a doctor, or the doctor who meets the matching degree Rm is unable to receive the patient, the consultation request is delivered to the order grabbing pool, and the doctor who meets the matching conditions takes the order actively, and displays the detailed information of the recommended doctor to the patient, including the professional background, historical consultation data and desensitized similar consultation cases for the patient to choose independently; At the same time, combined with the hospital scheduling system and the doctor's activity statistics by time period, evaluate the possibility of the doctor providing services currently, and set a matching degree adjustment factor Fa to affect the final sorting result. The calculation formula is as follows: ; Among them, is the adjusted matching degree; Fa is calculated by statistically analyzing the online rate, consultation situation and response time of the doctor in different time periods.

[0028] The triage optimization unit is used to calculate. Through the electronic medical record system, the cross-department matching index Ksp is obtained by counting the number of departments involved by the patient and the coverage range of the doctor's diagnosis and treatment departments. Through the hospital scheduling system, the current remaining available consultation slots of the doctor are obtained and analyzed in real time, and the doctor resource utilization rate Dzy is obtained by comparing with the maximum consultation capacity; Then, the demand matching degree Rm, the cross-department matching index Ksp and the doctor resource utilization rate Dzy are extracted and dimensionless processed, and the dynamic triage adjustment coefficient Fdz is calculated through the following formula: ; Read and based on the analysis of historical matching data, patient satisfaction feedback and doctor consultation load balancing strategy, preset the dynamic triage adjustment threshold F2, and compare and evaluate it with the dynamic triage adjustment coefficient Fdz. The specific evaluation content is as follows: If the dynamic triage adjustment coefficient Fdz ≥ the dynamic triage adjustment threshold F2, it means that the current matching plan meets the triage requirements, and the patient is matched to the current doctor at this time; If the dynamic triage adjustment coefficient Fdz < the dynamic triage adjustment threshold F2, it means that the current matching plan does not meet the triage requirements. At this time, dynamic triage adjustment includes: Adjust the cross-department matching tolerance Ksp to allow more doctors who meet part of the requirements to be included in the candidate list; Combined with the doctor resource utilization rate Dzy, guide the patient to be downward compatible according to the doctor recommendation order and match to a doctor with available consultation capacity; Finally, according to the doctor's real-time workload, including the number of online consultations and the conflict status of surgeries or offline consultations, multi-dimensional prompt information is generated, including: the response rate in the current period, the number of patients received during the same period, and the estimated waiting time; the system sets the minimum threshold of the professional matching degree Rm as the screening prerequisite condition, allowing patients to sort and screen candidate doctors according to multiple dimensions such as response probability and waiting time, and at the same time supporting doctors to set personalized consultation rules and consultation capacity limits through the maintenance interface. The privacy protection unit is used to desensitize the data related to patients' medical consultations, including de-identification, pseudo-anonymization, and encrypted data storage; first, by adopting a data classification and desensitization strategy, the desensitization methods for different levels of data are determined. For patients' identity information including names, ID numbers, and contact information, a hash function or irreversible encryption storage is used; for patients' medical records and diagnostic information, partial character masking is adopted, including character replacement for the medical record number and key fields in the medical record, while maintaining the integrity of the data format; for text information involving patients' privacy, including the main complaint description and past medical history.

[0029] In this embodiment, through the collaborative work of the triage optimization unit, matching calculation unit, and privacy protection unit of the doctor-patient matching module, the efficient matching and dynamic optimization of patients' needs and doctors' resources are realized; the matching calculation unit calculates the demand matching degree Rm based on the patient demand feature vector Fpd and the doctor feature vector Fys, and arranges the doctor recommendation order from high to low according to the matching degree Rm through the heap sorting algorithm to ensure that patients are matched to the most suitable doctor; the triage optimization unit calculates the cross-department matching index Ksp and the doctor resource utilization rate Dzy, generates a dynamic triage adjustment coefficient Fdz in combination with the demand matching degree Rm, and compares and evaluates it with the preset dynamic triage adjustment threshold F2, dynamically adjusting the matching plan to optimize the triage effect, including adjusting the cross-department matching tolerance Ksp, guiding patients to downwardly compatible match to doctors with spare consultation capacity, and recommending doctors with adjacent specialties or multi-disciplinary diagnosis and treatment capabilities, so as to improve the resource utilization efficiency and patient satisfaction; the privacy protection unit performs de-identification, pseudo-anonymization, and encrypted data storage on the data related to patients' medical consultations through the data classification and desensitization strategy to ensure the privacy and security of patients; the system significantly improves the matching accuracy, resource utilization efficiency, and patient privacy protection level through real-time calculation of the demand matching degree Rm, cross-department matching index Ksp, doctor resource utilization rate Dzy, and dynamic triage adjustment coefficient Fdz, and combined with the preset threshold F2 for dynamic adjustment, and finally realizes the efficient, accurate, and secure operation of the hospital companion diagnosis system.

[0030] In the calculation formula of the demand matching degree Rm, Fpd=(Fpd 1 , Fpd 2 ...Fpd n ), represents the patient demand feature vector, and each Fpdi is a characteristic value of the patient's needs, including Fpd 1 is the patient's disease type preference, Fpd 2 is the patient's treatment method preference, Fpd n is other need characteristics of the patient; Fys = (Fys 1 , Fys 2 ...Fys n ), representing the doctor feature vector, each Fys i is a characteristic value of the doctor, including Fys 1 is the disease type that the doctor is good at, Fys 2 is the treatment method that the doctor can provide, Fys n is other characteristics of the doctor; Among them, when calculating the Euclidean norm, square each eigenvalue, sum them up, and then take the square root to obtain the overall length of the feature vector; In addition, medical information is processed by differential privacy to prevent reverse derivation of identity. At the same time, secure transmission protocols including TLS and SSL are used for encrypted transmission to prevent man-in-the-middle attacks, and strict access control is carried out. Only the doctor-side system that meets the authorization policy is allowed to decrypt and obtain the necessary matching data; at the same time, before pushing the matching requirements, the system conducts a final review of the patient data to ensure that all sensitive information has been desensitized, and is pushed through compliant channels such as the internal secure data interface of the hospital or the certified doctor-patient matching platform to reduce the risk of information leakage; in addition, the matching information is transmitted through an encryption protocol and an access control policy to ensure that only authorized doctors can receive and decrypt. The doctor side can view the de-identified summary information and decide whether to accept the match based on the diagnosis and treatment ability and the reception status; the patient side can adjust or reject the match through an anonymous feedback channel to optimize the matching effect; the above solutions ensure patient privacy and security while achieving precise recommendation and dynamic triage adjustment, improving the utilization rate of doctor resources, and ensuring data security and compliance.

[0031] Example 4 The demand transformation module includes a resource scheduling unit and a demand evaluation and optimization unit; The resource scheduling unit first receives the patient's online consultation needs, analyzes the demand type, and identifies whether it involves face-to-face consultation, examination, or hospitalization arrangement; subsequently, by calling the API interface of the hospital business system, it queries and analyzes the current available status of medical resources, including doctor scheduling, inspection equipment idle time, and bed occupancy; after obtaining the medical resource status, the system attempts resource scheduling. During the scheduling execution process, the successful scheduling times Rac, the total scheduling attempt times Rat, and the resource availability rate Rar are recorded and collected. After dimensionless processing, the resource scheduling success rate Rds is calculated. The specific calculation formula is as follows: ; Based on the historical scheduling data of medical resources, calculate the optimal scheduling success rate threshold under different medical scenarios through statistical analysis and regression analysis machine learning models. Preset the resource scheduling success rate threshold R, and compare and evaluate it with the resource scheduling success rate Rds. The specific evaluation content is as follows: The resource scheduling success rate Rds ≥ the resource scheduling success rate threshold R indicates that the resource scheduling is successful, the current medical resources meet the requirements, and the system executes business processes such as face-to-face consultations, examinations, and hospitalization arrangements according to the established demand allocation strategy; at the same time, enter the calculation link of the demand conversion rate Ctr to evaluate the proportion of online demands successfully converted into offline consultations. The resource scheduling success rate Rds < the resource scheduling success rate threshold R indicates that the resource scheduling is unsuccessful, and the current medical resources do not meet the requirements. The reasons include that due to overbooking of doctor appointments, tight examination schedules, or insufficient hospital beds, some demands cannot be scheduled; at this time, the system needs to adjust the demand allocation strategy, including: re-matching available resources, guiding patients to other medical institutions or departments; And only after the adjustment, re-evaluate the resource scheduling success rate Rds until it is greater than or equal to the resource scheduling success rate threshold R, the system will enter the calculation and evaluation link of the demand conversion rate Ctr, otherwise continue to adjust the scheduling strategy.

[0032] The demand evaluation and optimization unit is responsible for calculating and evaluating the demand conversion rate Ctr; based on the actual consultation records of patients in the hospital business system, including the proportion statistics of the number of reserved face-to-face consultations, the number of completed examinations, and the number of hospitalization arrangements in the online consultation demands, obtain the face-to-face consultation proportion Cmz, the examination proportion Cjc, and the hospitalization arrangement proportion Czh; After extracting the face-to-face consultation proportion Cmz, the examination proportion Cjc, and the hospitalization arrangement proportion Czh and performing dimensionless processing, calculate and obtain the demand conversion rate Ctr through the following formula:

[0033] Based on the historical conversion data, calculate the lowest acceptable conversion rate under different medical scenarios through statistical analysis and clustering analysis machine learning models. Preset the demand conversion rate threshold C, and compare and evaluate it with the demand conversion rate Ctr. The specific evaluation content is as follows: The demand conversion rate Ctr ≥ the demand conversion rate threshold C indicates that the demand conversion rate is qualified, the conversion process is normal, and the proportion of online consultations successfully converted into offline services reaches or exceeds the preset threshold, indicating that the accompanying consultation process, resource scheduling, and patient demand matching of the system are all at a normal level; When the demand conversion rate Ctr < the demand conversion rate threshold C, it indicates that the demand conversion rate is unqualified and the conversion process is abnormal. The reasons for the unqualified demand conversion rate include factors such as the failure to complete the appointment smoothly, the patient's abandonment of the face-to-face consultation, the conflict in the examination schedule, or the tightness of the hospital beds; at this time, the system needs to adjust the demand conversion process, including: adjusting the matching strategy, improving the appointment mechanism, and adjusting the patient guidance. After the optimization and adjustment, re-evaluate the demand conversion rate Ctr until the demand conversion rate is qualified.

[0034] In this embodiment, through the collaborative work of the resource scheduling unit and the demand assessment and optimization unit of the demand conversion module, the efficient conversion and dynamic optimization of online consultation demands into offline services are realized; the resource scheduling unit parses the patient demand type and calls the hospital business system API interface to query the medical resource status, calculates the resource scheduling success rate Rds, and compares and evaluates it with the preset resource scheduling success rate threshold R, and dynamically adjusts the demand allocation strategy to ensure the success of resource scheduling, including re-matching available resources or guiding patients to other medical institutions; the demand assessment and optimization unit calculates the demand conversion rate Ctr by counting the face-to-face consultation ratio Cmz, the examination ratio Cjc, and the hospitalization arrangement ratio Czh, and compares and evaluates it with the preset demand conversion rate threshold C, and dynamically optimizes the demand conversion process to improve the conversion efficiency, including adjusting the matching strategy, improving the appointment mechanism, and optimizing the patient guidance; the system dynamically adjusts by real-time collecting and calculating the resource scheduling success rate Rds and the demand conversion rate Ctr, and combining the preset thresholds R and the demand conversion rate threshold C, significantly improves the resource scheduling efficiency, the demand conversion success rate, and the patient's medical experience, and finally realizes the efficient connection between online consultation and offline services and the optimized allocation of hospital resources.

[0035] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A companion diagnosis system based on online consultation service, characterized in that: It includes patient demand collection module, demand classification module, rule filtering module, doctor feature vector library module, doctor-patient matching module and correction and adjustment module; The patient demand collection module collects patient consultation data through the mobile terminal and uses natural language processing to extract the patient demand feature vector Fpd; The composite demand identification and decomposition module receives the patient demand feature vector Fpd and identifies whether it is a composite demand. The patient demand feature vector Fpd includes information inquiry, transactional demand, health consultation and medical consultation; When the demand feature vector Fpd contains at least two demands, it is identified as a composite demand, and the large language model LLM technology is used to decompose the patient demand feature vector Fpd into multiple independent single feature demands; The rule filtering module is used to build in predefined rules, including pediatric age restrictions, gender department restrictions, and emergency priority, to filter out unreasonable matches; Doctor feature vector library module, used to construct doctor feature vectors Fys of the professional category of the selected doctor and service feature indicators, where the doctor feature vector Fys is generated through topic modeling of historical valid consultation data; The doctor-patient matching module is used to calculate the matching degree Rm between each independent feature demand in the patient demand feature vector Fpd and the doctor feature vector Fys, and rank and recommend them; at the same time, the consultation request is delivered to the order grabbing pool, and the doctor who meets the matching conditions will take the initiative to accept the order; The correction and adjustment module is used to evaluate the resource scheduling success rate Rds, adjust the demand conversion process, and provide an interface for doctors to maintain the feature corpus.

2. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: The demand classification module includes a demand identification unit and a demand association analysis unit; After receiving the medical consultation-related data, the demand identification unit extracts key information from the text by performing text parsing, including word segmentation, part-of-speech tagging, syntactic analysis and entity recognition; then, based on the deep learning neural network, a semantic classification model is constructed and trained, and the medical consultation-related data is classified and processed through the semantic classification model, and the semantic similarity Fy between the input text and the demand category is calculated. Finally, information query, transactional needs, health consultation and medical consultation are identified, and different demand categories are split based on semantic similarity and association rule mining. By analyzing historical patient consultation data, the co-occurrence relationship of different categories of needs is counted, the demand co-occurrence probability Fg is obtained, and the corresponding demand category label is constructed; finally, by tracking the contextual information of the patient's medical consultation text, the changes in the content input by the patient before and after are analyzed, and the context relevance Fs is obtained.

3. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: The demand association analysis unit is used to analyze the logical relationship between various demand categories and make demand adjustments and matching adjustments; Through semantic similarity analysis, co-occurrence relationship mining and context association analysis, the consultation data of patients in online consultation is obtained, and the consultation data is semantically analyzed and classified to obtain the consultation classification results. The consultation classification results are matched and analyzed with the pre-processed historical patient consultation data to calculate the semantic similarity Fy, demand co-occurrence probability Fg and context correlation Fs respectively. After dimensionless processing and fitting, the correlation Fgx of the composite demand is calculated. The specific calculation formula is as follows: ； Based on the historical patient demand classification data, the average correlation of different demand categories is extracted through statistical analysis and decision tree machine learning model, and the correlation threshold F1 is preset; the correlation threshold F1 is compared and evaluated with the correlation Fgx to make demand adjustments and matching adjustments; the specific evaluation contents are as follows: When the correlation Fgx ≥ the correlation threshold F1, it is determined that the complex needs in the current medical consultation data do not need to be split, and the integrity is retained. In the matching process, comprehensive doctors or departments containing multiple categories are recommended; When the correlation degree Fgx is less than the correlation threshold F1, it is determined that the complex needs in the current medical consultation-related data need to be split, and each need is matched to the corresponding doctor or department, and personalized recommendation adjustments are made to the patient's needs.

4. A companion diagnosis system based on online consultation service according to claim 3, characterized in that: The rule filtering module intelligently screens patient needs based on the preset medical business rule library. First, an age recognition mechanism is established to analyze the patient's age information. When it is detected that the patient is under 14 years old and the disease type does not belong to the trauma category, the pediatric allocation rule is automatically triggered to direct the current patient's needs to the pediatric diagnosis and treatment queue. Secondly, a gender and department matching mechanism is established and implemented to cross-verify the patient's gender information and the department's scope of treatment. When a male patient is identified, the matching path of obstetricians and gynecologists is automatically blocked; Finally, an emergency classification mechanism is established to prioritize the patient's condition. When a patient is identified as a critically ill patient, the priority matching mechanism is automatically activated to push the patient's needs to the emergency treatment channel of the corresponding specialist. In addition, the age recognition mechanism, gender and department matching mechanism, and emergency classification mechanism in the rule filtering module realize compound condition judgment through logical AND operations.

5. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: During operation, the doctor feature vector library module adopts a dual-path matching mechanism. On the one hand, it performs topic modeling on the desensitized historical consultation data based on the LDA algorithm, constructs the doctor's topic feature vector, and counts the service indicators; on the other hand, it stores the patient consultation corpus in a feature-based manner, retrieves and matches the most relevant N historical corpora through similarity when a new consultation is made, and recommends doctors based on the distribution of doctors in the corpus through a voting mechanism; At the same time, a doctor maintenance interface is set up for doctors to regularly review and correct feature data. Finally, recommendations are generated by weighted fusion of the matching results of the topic distribution path and the corpus similarity path. The weight of the topic path increases with the accumulation of corpus, while the weight of the similarity path is kept at a high ratio at the beginning of a new topic to ensure matching accuracy. At the same time, service indicators including doctor's patient satisfaction, consultation response rate and response timeliness are counted in real time; in addition, a doctor maintenance interface is set up to allow doctors to regularly review and revise the topic classification in the feature corpus, so as to adjust the professionalism and timeliness of the feature vector, and finally output comprehensive doctor feature data including multi-dimensional professional feature vectors and quantitative service indicators.

6. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: The doctor-patient matching module includes a matching calculation unit, a triage optimization unit, and a privacy protection unit; The matching calculation unit is used to calculate the matching degree Rm of each patient's needs based on the patient's needs feature vector Fpd and the doctor's feature vector Fys; first, the matching degree Rm of the patient's needs is defined as the specific calculation formula as follows: ； In the formula, represents the dot product of the patient demand feature vector and the doctor feature vector; represents the Euclidean norm of the patient demand feature vector, and the patient demand feature vector Fpd is composed of multiple demand-related features; Finally, the heap sort algorithm is used to sort the doctor recommendations from high to low according to the value of the matching degree Rm. After the sorting is completed, the system recommends the doctor with the highest matching degree based on the sorting results. When the patient has not specified a doctor, or the doctor who meets the matching degree Rm is unable to see the patient, the consultation request will be delivered to the order grabbing pool, and the doctor who meets the matching conditions will take the initiative to accept the order. The detailed information of the recommended doctor, including professional background, historical consultation data and similar consultation cases that have been desensitized, will be displayed to the patient for the patient to choose. At the same time, combined with the hospital scheduling system and the statistics of doctors' activity in different time periods, the possibility of doctors providing services at present is evaluated, and the matching adjustment factor Fa is set to affect the final ranking result. The calculation formula is as follows: ； in, is the adjusted matching degree; Fa is calculated by counting the online rate, consultation status and response time of doctors in different time periods.

7. A companion diagnosis system based on online consultation service according to claim 6, characterized in that: The triage optimization unit is used to calculate, through the electronic medical record system, the number of departments involved by the patient and the coverage of the doctor's diagnosis and treatment departments to obtain the cross-department matching index Ksp, and through the hospital scheduling system, obtain and analyze the current remaining doctor's available appointments in real time and obtain the doctor's resource utilization rate Dzy with the maximum appointment capacity; Then, the demand matching degree Rm, the cross-department matching index Ksp and the doctor resource utilization rate Dzy are extracted and dimensionless processed, and the dynamic triage adjustment coefficient Fdz is calculated by the following formula: ； The dynamic triage adjustment threshold F2 is read and preset based on historical matching data analysis, patient satisfaction feedback, and doctor reception load balancing strategy, and compared and evaluated with the dynamic triage adjustment coefficient Fdz. The specific evaluation contents are as follows: If the dynamic triage adjustment coefficient Fdz ≥ the dynamic triage adjustment threshold F2, it means that the current matching scheme meets the triage requirements, and the patient is matched to the current doctor at this time; If the dynamic triage adjustment coefficient Fdz is less than the dynamic triage adjustment threshold F2, it means that the current matching scheme does not meet the triage requirements. In this case, the dynamic triage adjustment includes: Adjust the cross-department matching tolerance Ksp to allow more doctors who meet some requirements to be included in the candidate list; Combined with the doctor resource utilization rate Dzy, patients are guided to be downward compatible according to the doctor recommendation order and matched with doctors with spare consultation capacity; Finally, based on the doctor's real-time workload, including the number of online consultations, and the conflict status of surgery or offline outpatient clinics, multi-dimensional prompt information is generated, including: the response rate of the current period, the number of patients seen during the same period, and the estimated waiting time; the system sets the minimum threshold of the professional matching degree Rm as a screening prerequisite, allowing patients to sort and screen candidate doctors based on multiple dimensions such as response probability and waiting time, and at the same time supports doctors to set personalized consultation rules and reception capacity limits through the maintenance interface.

8. A companion diagnosis system based on online consultation service according to claim 6, characterized in that: The privacy protection unit is used to desensitize patient medical data, including de-identification, pseudo-anonymization and encrypted data storage; first, by adopting a hierarchical data desensitization strategy, the desensitization methods for data of different levels are determined. For patient identity information including patient name, ID number and contact information, hash functions or irreversible encryption storage are used; for patient medical records and diagnostic information, partial character masking is used, including character replacement of key fields in medical record numbers and medical records, while maintaining the integrity of the data format; for text information involving patient privacy, including chief complaint description and past medical history.

9. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: The demand conversion module includes a resource scheduling unit and a demand assessment and optimization unit; The resource scheduling unit first receives the patient's online consultation needs, analyzes the demand type, and identifies whether it involves face-to-face consultation, examination, or hospitalization arrangements; then, by calling the hospital business system API interface, it queries and analyzes the current available status of medical resources, including doctor scheduling, examination equipment idle time, and bed occupancy; after obtaining the medical resource status, the system executes resource scheduling attempts. During the scheduling execution process, the number of successful scheduling attempts Rac, the total number of scheduling attempts Rat, and the resource availability rate Rar are recorded and collected. After dimensionless processing, the resource scheduling success rate Rds is calculated. The specific calculation formula is as follows: ； Based on the historical scheduling data of medical resources, the optimal scheduling success rate threshold in different medical scenarios is calculated through statistical analysis and regression analysis machine learning models. The preset resource scheduling success rate threshold R is compared with the resource scheduling success rate Rds for evaluation. The specific evaluation contents are as follows: If the resource scheduling success rate Rds ≥ the resource scheduling success rate threshold R, it means that the resource scheduling is successful, the current medical resources meet the demand, and the system executes the business processes such as face-to-face consultation, examination and hospitalization arrangement according to the established demand allocation strategy; at the same time, it enters the calculation link of the demand conversion rate Ctr to evaluate the proportion of online demand successfully converted into offline medical treatment; If the resource scheduling success rate Rds is less than the resource scheduling success rate threshold R, it means that the resource scheduling is unsuccessful and the current medical resources do not meet the demand. The reasons include that some demands cannot be scheduled due to overbooked doctors, tight examination schedules or insufficient hospital beds. At this time, the system needs to adjust the demand allocation strategy, including: re-matching available resources and guiding patients to other medical institutions or departments; And only after the adjustment, the resource scheduling success rate Rds is re-evaluated until it is greater than or equal to the resource scheduling success rate threshold R. The system then enters the calculation and evaluation phase of the demand conversion rate Ctr, otherwise it continues to adjust the scheduling strategy.

10. A companion diagnosis system based on online consultation service according to claim 1, characterized in that: The demand assessment and optimization unit is responsible for calculating and evaluating the demand conversion rate Ctr; based on the actual medical records of patients in the hospital business system, including the number of scheduled face-to-face consultations, the number of completed examinations, and the proportion of hospitalization arrangements to online consultation needs, the face-to-face consultation ratio Cmz, the examination ratio Cjc, and the hospitalization ratio Czh are obtained; After extracting the face-to-face consultation ratio Cmz, the examination ratio Cjc, and the hospitalization arrangement ratio Czh and performing dimensionless processing, the demand conversion rate Ctr is calculated using the following formula: Based on historical conversion data, the lowest acceptable conversion rate in different medical scenarios is calculated through statistical analysis and cluster analysis machine learning models. The demand conversion rate threshold C is preset and compared with the demand conversion rate Ctr for evaluation. The specific evaluation contents are as follows: Demand conversion rate Ctr≥demand conversion rate threshold C, indicating that the demand conversion rate is qualified, the conversion process is normal, and the proportion of online consultation successfully converted into offline business reaches or exceeds the preset threshold, indicating that the system's accompanying diagnosis process, resource scheduling and patient demand matching are all at a normal level; Demand conversion rate Ctr < demand conversion rate threshold C, indicating that the demand conversion rate is unqualified and the conversion process is abnormal. The reasons for the failure to meet the demand conversion rate include failure to complete the appointment smoothly, patients giving up face-to-face consultation, examination scheduling conflicts, or shortage of hospital beds; At this time, the system needs to adjust the demand conversion process, including: adjusting the matching strategy, improving the appointment mechanism, and adjusting patient guidance; After optimization and adjustment, re-evaluate the demand conversion rate Ctr until the demand conversion rate is qualified.

Images