A multi-terminal adaptive labeling system for preoperative information entry for children

Abstract

The present application relates to the technical field of information entry, and discloses a multi-terminal adaptive labeling system for preoperative information entry of children, which comprises a multi-terminal interaction module for receiving preoperative basic information of a single child from various terminals; an integrity checking module for checking whether the preoperative basic information entered by any terminal is missing, and if so, generating a corresponding preoperative core information dimension-specific completion prompt until the preoperative complete information of the single child is obtained; an adaptive labeling module for determining an adaptive labeling granularity based on the type of the terminal for entering the preoperative complete information, and dynamically matching and labeling non-professional entry information in the preoperative complete information of the single child based on a child specialty terminology library and the adaptive labeling granularity; and a secondary verification module for triggering a secondary verification prompt based on structured evaluation items and cross-terminal entry conflict information until the preoperative standard information is obtained; thus, the quality of preoperative information of children is ensured, and reliable basis is provided for preoperative evaluation of children.

Description

Technical Field

[0001] This invention relates to the field of information entry technology, and in particular to a multi-terminal adaptive annotation system for preoperative information entry in children. Background Technology

[0002] In the medical field, accurate entry of preoperative information for children is crucial for the smooth conduct of surgery and the safety of patients. With the development of digital technology, more and more medical information management is being accomplished using various terminal devices, making multi-terminal entry of children's preoperative information a trend. However, the situation regarding information entry across different terminals is complex, posing numerous challenges to ensuring the completeness, accuracy, and standardization of the information. Children, as a special medical group, have preoperative information involving multiple core dimensions, including but not limited to basic health status, past medical history, and allergy history. The comprehensive and accurate collection of this information plays a decisive role in the formulation of surgical plans. In today's fast-paced medical environment, efficient and accurate entry of children's preoperative information helps improve medical efficiency, reduce surgical risks, and enhance the overall quality of medical services. Simultaneously, with the continuous advancement of internet and mobile technologies, the application of multi-terminal devices in medical scenarios is becoming increasingly widespread. For example, hospital self-service terminals, mobile tablets used by medical staff, and mobile phones used by parents or guardians can all serve as entry points for information. Developing a preoperative information entry and annotation system for children that can adapt to multiple terminals can not only meet the information entry needs in different scenarios, but also make full use of modern technology to improve the level of medical informatization, lay the foundation for the future development of smart healthcare, and has broad application prospects.

[0003] However, existing methods for entering preoperative information into pediatric hospitals lack effective mechanisms to ensure the completeness of the entered information when inputting data across multiple devices. This leads to the omission of crucial information and affects the accuracy of surgical planning. For information entered from different devices, it is impossible to determine the appropriate annotation granularity based on the device type, and it is difficult to dynamically match and annotate non-specialist information using a pediatric terminology database, resulting in insufficient standardization and professionalism of the information. Furthermore, the existing methods lack mechanisms to trigger secondary verification prompts, failing to guarantee that the final preoperative information meets the required standards, potentially posing risks to the surgery.

[0004] Therefore, this invention proposes a multi-terminal adaptive annotation system for preoperative information entry in children. Summary of the Invention

[0005] This invention provides a multi-terminal adaptive annotation system for preoperative information entry in children. It can receive basic preoperative information from a single child from various compatible terminals, facilitating information collection and meeting information entry needs in different scenarios. Based on the mandatory field verification rules for all core preoperative information dimensions, it detects missing information and generates targeted completion prompts, ensuring the acquisition of complete preoperative information that passes verification, thus improving information integrity and accuracy. The annotation granularity is determined according to the type of input terminal. A pediatric terminology database is used to dynamically match and annotate non-professional input information, resulting in preoperative annotation information for a single child. This makes the annotation more relevant to the actual input situation, improving the professionalism and standardization of the information. Based on the preoperative annotation information and a preoperative assessment standard library for children, structured assessment items are generated. It identifies conflicting information entered across terminals and triggers secondary verification prompts until standard preoperative information is obtained, further ensuring information quality and providing a reliable basis for preoperative assessment. This helps improve the efficiency and quality of preoperative preparation and better ensures the smooth progress of pediatric surgery.

[0006] This invention provides a multi-terminal adaptive annotation system for preoperative information entry in children, comprising:

[0007] A multi-terminal interaction module is used to receive preoperative basic information of a single child from multiple pre-adapted terminal sources;

[0008] The integrity verification module is used to detect whether there are any missing preoperative basic information entered by any terminal based on the verification rules of the required fields of all preoperative core information dimensions. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions until the complete preoperative information of a single child that has passed the verification is obtained.

[0009] The adaptive annotation module is used to determine the adaptive annotation granularity based on the type of input terminal for complete preoperative information, and to dynamically match and annotate non-professional input information in the complete preoperative information of a single child based on the pediatric terminology database and the adaptive annotation granularity, so as to obtain the preoperative annotation information of a single child.

[0010] The secondary verification module is used to associate the preoperative assessment standard library for children with the preoperative annotation information and generate structured assessment items. It identifies cross-terminal input conflict information in the preoperative annotation information, triggers secondary verification prompts based on the structured assessment items and cross-terminal input conflict information, until the preoperative standard information is obtained.

[0011] Preferably, the multi-terminal interaction module includes:

[0012] The hospital management information system's interactive submodule is used to call the structured data interface to directly read the basic fields of a single child's electronic medical record and obtain preoperative basic information from the hospital management information system.

[0013] The medical mobile terminal interaction submodule is used to input basic preoperative information of a single child from the medical mobile terminal based on a preset touch form;

[0014] The parent self-service terminal interaction submodule is used to input basic preoperative information of a single child from the parent self-service terminal based on the icon-based selection command entry of the preset graphic interaction interface.

[0015] Preferably, the integrity verification module includes:

[0016] The validation rule setting submodule is used to set the validation rules for the required fields of each preoperative core information dimension based on the child's age stratum.

[0017] The integrity verification submodule is used to detect whether there are any missing preoperative basic information entered by any terminal based on the mandatory field verification rules. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions based on all preoperative core information dimensions of children that do not have any missing information, until the complete preoperative information of a single child that has passed the verification is obtained.

[0018] Preferably, the adaptive annotation module includes:

[0019] The adaptive annotation granularity determination submodule is used to determine the adaptive annotation granularity based on the type of terminal used to input complete preoperative information.

[0020] The symptom entity recognition submodule is used to perform symptom entity recognition on the complete preoperative information of a single child to obtain all symptom entities of the corresponding child.

[0021] The terminology graph retrieval submodule is used to generate a pediatric terminology knowledge graph based on a pediatric terminology database, and to determine the matching standard terms for all non-professional input information of all symptom entities of a single child based on the annotation granularity and the pediatric terminology knowledge graph.

[0022] The matching information standard submodule is used to annotate all non-professional input information of a single child with matching standard terms and then to all non-professional input information in the corresponding child's complete preoperative information to obtain the corresponding child's preoperative annotation information.

[0023] Preferably, the terminology map retrieval submodule includes:

[0024] The knowledge graph building unit is used to determine all professional and non-professional descriptions of all symptom terminology entities based on the pediatric specialist terminology database. Based on the correlation coefficients between all symptom entities involved in the pediatric specialist terminology database and all professional and non-professional descriptions of all symptom terminology entities, a pediatric specialist terminology knowledge graph is generated.

[0025] The matching degree calculation unit is used to determine the matching degree between each symptom entity and each specialty symptom term entity based on all non-professional and professional input information of all symptom entities and all non-professional and professional description information in the pediatric specialty terminology knowledge graph.

[0026] The centralized positioning unit is used to perform centralized positioning in the pediatric specialty terminology knowledge graph based on the matching degree between each symptom entity and each specialty symptom term entity, and to determine the current centralized location range in the pediatric specialty terminology knowledge graph.

[0027] The professional retrieval depth determination unit is used to determine the professional retrieval depth based on the current concentration range and adaptive annotation granularity in the pediatric specialty terminology knowledge graph.

[0028] The standard term matching unit is used to retrieve matching standard terms for all non-professional input information of all symptom entities in the specialist symptom standard term conversion database based on professional retrieval depth.

[0029] Preferably, the matching degree calculation unit includes:

[0030] Based on the semantic similarity between each non-professional entry information of each symptom entity and each non-professional description information of each specialty symptom term entity, and the semantic similarity between each professional entry information of each symptom entity and each professional description information of each specialty symptom term entity;

[0031] Based on the semantic similarity between all non-professional input information of each symptom entity and each non-professional description information of each specialist symptom term entity, and the semantic similarity between all professional input information of the corresponding symptom entity and each professional description information of the corresponding specialist symptom term entity, the matching degree between each symptom entity and the corresponding specialist symptom term entity is calculated.

[0032] Preferably, the centralized positioning unit includes:

[0033] The high-match entity filtering subunit is used to filter out all specialist symptom terminology entities in the pediatric specialist terminology knowledge graph whose match degree with each symptom entity exceeds a preset match degree threshold as all high-match entities of each symptom entity.

[0034] Temporary subnetwork localization subunits are used to identify temporary association subnetworks in the pediatric terminology knowledge graph based on the direct association paths between all highly matched entities of all symptom entities.

[0035] The first centralized location subunit is used to treat the temporary associated subnetwork as the current centralized location range in the children's specialty terminology knowledge graph when the average association coefficient of the temporary associated subnetwork is greater than the first association coefficient threshold and the node clustering coefficient is greater than the second association threshold.

[0036] The second centralized location subunit is used to perform a stepwise increase in the matching degree when the average correlation coefficient of the temporary associated subnetwork is not greater than the first correlation coefficient threshold or the node clustering coefficient is not greater than the second correlation coefficient threshold. Based on the latest matching value, all high-matching entities of each symptom entity are re-filtered until the average correlation coefficient of the temporary associated subnetwork determined based on all high-matching entities of all the latest filtered symptom entities is greater than the first correlation coefficient threshold and the node clustering coefficient is greater than the second correlation coefficient threshold. Then, the newly determined temporary associated subnetwork is regarded as the current centralized location range in the pediatric specialty terminology knowledge graph.

[0037] Preferred, specialized search depth determination units include:

[0038] The core term entity selection sub-unit is used to determine the comprehensive matching degree between each specialty symptom term entity and all symptom entities in the current centralized location range, and the specialty symptom term entity with the largest comprehensive matching degree in the current centralized location range is regarded as the core term entity.

[0039] Multi-level term entity extension subunits are used to identify multiple levels of extended entity groups from the current centralized location range, centered on the core term entity and in descending order of the correlation coefficient.

[0040] The location vector generation subunit is used to generate the location vector of each level of the extended entity group based on the distance between all edge nodes of each level of the extended entity group and all edge nodes of the children's specialty terminology knowledge graph.

[0041] The retrieval depth determination subunit is used to determine the specialized retrieval depth based on the location vectors and adaptive annotation granularity of the extended entity group at all levels.

[0042] Preferably, the retrieval depth determination subunit includes:

[0043] The first retrieval depth determination end is used to determine the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth. Based on the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth and the positioning vector of the extended entity group at each level, the first sub-professional retrieval depth corresponding to the extended entity group at each level is calculated.

[0044] The second retrieval depth determination end is used to determine the second sub-specialty retrieval depth based on the adaptive annotation granularity and the annotation granularity-specialty retrieval depth mapping rule;

[0045] The third search depth determination end is used to determine the professional search depth based on the search depths of all first sub-professions and second sub-professions.

[0046] Preferably, the secondary verification module includes:

[0047] The assessment item generation submodule is used to automatically generate structured assessment items by associating with the preoperative assessment standard library for children based on preoperative annotation information;

[0048] The conflict identification submodule is used to identify cross-terminal input conflict information in the preoperative annotation information and associate conflicting terminals;

[0049] The verification path determination submodule is used to determine the verification path based on structured evaluation items, cross-terminal input conflict information, and conflicting terminals.

[0050] The verification prompt submodule is used to trigger secondary verification prompts based on the verification path until the preoperative standard information is obtained.

[0051] The beneficial effects of this invention compared to existing technologies are as follows: It can receive basic preoperative information of a single child from multiple compatible terminals, facilitating information collection and meeting the information entry needs in different scenarios. Based on the mandatory field verification rules for all core preoperative information dimensions, it detects missing information and generates targeted completion prompts, ensuring the acquisition of complete preoperative information that passes verification, thus improving information integrity and accuracy. The annotation granularity is determined according to the type of input terminal, and non-professional input information is dynamically matched and annotated using a pediatric terminology database to obtain preoperative annotation information for a single child, making the annotation more consistent with the actual input situation and improving the professionalism and standardization of the information. Based on the preoperative annotation information, a pediatric preoperative assessment standard library is associated to generate structured assessment items, identify conflicting information entered across terminals and trigger secondary verification prompts until standard preoperative information is obtained, further ensuring information quality, providing a reliable basis for preoperative assessment of children, helping to improve the efficiency and quality of preoperative preparation, and better ensuring the smooth progress of pediatric surgery.

[0052] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0053] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0055] Figure 1This is a schematic diagram of a multi-terminal adaptive annotation system for recording preoperative information of children in an embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram of the multi-terminal interaction module in an embodiment of the present invention;

[0057] Figure 3 This is a schematic diagram of the integrity verification module in an embodiment of the present invention;

[0058] Figure 4 This is a schematic diagram of the adaptive annotation module in an embodiment of the present invention. Detailed Implementation

[0059] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0060] like Figure 1 As shown, this invention provides an implementation method for a multi-terminal adaptive annotation system for preoperative information entry in children, comprising:

[0061] A multi-terminal interaction module is used to receive preoperative basic information of a single child from multiple pre-adapted terminal sources;

[0062] The integrity verification module is used to detect whether there are any missing preoperative basic information entered by any terminal based on the verification rules of the required fields of all preoperative core information dimensions. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions until the complete preoperative information of a single child that has passed the verification is obtained.

[0063] The adaptive annotation module is used to determine the adaptive annotation granularity based on the type of input terminal for complete preoperative information, and to dynamically match and annotate non-professional input information in the complete preoperative information of a single child based on the pediatric terminology database and the adaptive annotation granularity, so as to obtain the preoperative annotation information of a single child.

[0064] The secondary verification module is used to associate the preoperative assessment standard library for children with the preoperative annotation information and generate structured assessment items. It identifies cross-terminal input conflict information in the preoperative annotation information, triggers secondary verification prompts based on the structured assessment items and cross-terminal input conflict information, until the preoperative standard information is obtained.

[0065] In this embodiment, the pre-adapted multiple terminal sources refer to various types of terminal devices that the system has adapted in advance to receive information.

[0066] In this embodiment, preoperative basic information refers to the basic data related to the surgery that a single child enters from various compatible terminals before the surgery. This includes aspects such as basic health status, past medical history, and allergy history.

[0067] In this embodiment, the core preoperative information dimension is a key category of information classification within a child's preoperative information, which is extremely important for comprehensively understanding the child's preoperative condition and formulating a surgical plan. For example, basic health status, past medical history, and allergy history all belong to different core preoperative information dimensions, and each dimension contains a series of important information.

[0068] In this embodiment, the mandatory field verification rules for the preoperative core information dimensions are as follows: based on the child's age group, rules are set for each preoperative core information dimension to determine whether specific information under that dimension is mandatory. For example, in the past medical history dimension for children of a certain age group, "whether they have had a specific serious illness" may be set as a mandatory field, thereby ensuring that no key information is omitted and providing a comprehensive and accurate basis for surgical planning.

[0069] In this embodiment, the system checks for missing preoperative basic information entered by any terminal based on the mandatory field validation rules for all core preoperative information dimensions. The system checks the preoperative basic information entered from any terminal, such as the hospital management information system, medical staff mobile terminals, or parent self-service terminals, according to the mandatory field validation rules defined for each core preoperative information dimension. For example, if the allergy history dimension specifies "specific allergen" as a mandatory field, and this field is blank in the information entered by a certain terminal, it is determined that the preoperative basic information entered by this terminal is missing in the allergy history core information dimension.

[0070] In this embodiment, targeted completion prompts are provided for preoperative core information dimensions: When a missing preoperative basic information dimension is detected, the system generates a specific prompt message for that missing dimension based on existing preoperative sub-basic information for the child in the missing preoperative core information dimension, guiding the information entry user to complete the information in that dimension. For example, if "height" information is detected as missing in the basic health status dimension, the system may prompt, "Please complete the child's height information, as this is important for a comprehensive assessment of the child's health status."

[0071] In this embodiment, the complete preoperative information of a single child that passes verification refers to the set of information that, after being checked by the integrity verification module based on the mandatory field verification rules of all core preoperative information dimensions, has no missing information and meets the requirements for the comprehensiveness of preoperative child information. Only information that meets this standard can be used for subsequent adaptive annotation and other operations, providing a reliable basis for surgical planning.

[0072] In this embodiment, the terminal type for entering complete preoperative information refers to the type of terminal used to enter verified complete preoperative information, namely, one of the following: hospital management information system terminal, medical staff mobile terminal, or parent self-service terminal. Different terminal types may reflect differences in information entry scenarios, the identity of the person entering the information, etc., and the system will determine the adaptive annotation granularity accordingly.

[0073] In this embodiment, the adaptive annotation granularity is determined based on the type of terminal used to input complete preoperative information: the system determines the level of detail and accuracy of information annotation based on whether the terminal used to input complete preoperative information is the hospital management information system, a medical staff mobile terminal, or a parent self-service terminal. For example, information entered through a parent self-service terminal may contain more non-professional descriptions, so the annotation granularity may be finer to better match non-professional information with professional terminology; while information entered through the hospital management information system may be more standardized, so the annotation granularity is relatively coarser.

[0074] In this embodiment, the adaptive annotation granularity refers to the level of detail and accuracy of annotation on non-professional information, adjusted according to the type of terminal used for preoperative complete information entry. Different annotation granularities determine the level of detail the system achieves when dynamically matching and annotating non-professional information using a pediatric terminology database, making the annotations more tailored to the characteristics of information entered from different terminals and improving the professionalism and standardization of the information.

[0075] In this embodiment, the pediatric specialist terminology database is a database containing professional terms in the field of pediatric medicine and related information. It stores professional and non-professional descriptions of all specialist symptom terminology entities, as well as the correlation coefficients between various symptom entities.

[0076] In this embodiment, non-professional input information refers to information expressed in ordinary, non-medical language within the complete preoperative information. For example, parents may describe their child's symptoms using everyday language, which differs from medical terminology. This constitutes non-professional input information and requires dynamic matching and annotation by the system based on a pediatric terminology database and adaptive annotation granularity to transform it into professional and standardized expressions.

[0077] In this embodiment, the preoperative annotation information refers to non-professional input information within the complete preoperative information. This information is obtained by dynamically matching and annotating non-professional expressions with professional terminology using an adaptive annotation granularity determined by the input terminal type, and utilizing a pediatric specialist terminology database. In other words, it's pediatric preoperative information annotated with professional terms, making the information more standardized and professional, facilitating subsequent operations such as linking it to a pediatric preoperative assessment standard database.

[0078] In this embodiment, the pediatric preoperative assessment standard library is a database that stores the standard information required for pediatric preoperative assessment, including the standards and specifications upon which various preoperative assessments are based. The system associates the preoperative annotation information with this standard library to generate structured assessment items, providing a basis for determining whether the preoperative information meets the standards and for secondary verification.

[0079] In this embodiment, the structured assessment items are assessment entries with a certain structure and standardization, generated based on the preoperative annotation information and linked to a pediatric preoperative assessment standard library. These may include assessment standards for various health indicators of children, surgical risk assessment items, etc., presented in a structured form to facilitate system analysis and judgment of whether preoperative information meets the standards, as well as to identify conflicting information entered across terminals.

[0080] In this embodiment, conflicting information entered across terminals refers to contradictory or inconsistent content in preoperative annotation information entered from different terminals. For example, if the child's allergy history entered into the hospital management information system does not match the allergy history information entered into the parent's self-service terminal, this discrepancy constitutes conflicting information entered across terminals. The system will trigger a secondary verification prompt upon recognizing such information.

[0081] In this embodiment, a secondary verification prompt is triggered based on structured assessment items and cross-terminal input conflict information: the system sends a prompt to relevant personnel to re-verify the information based on the generated structured assessment items and the identified cross-terminal input conflict information. For example, when the structured assessment items show that a certain indicator does not meet the standard, and there is also cross-terminal input conflict information, the system prompts medical staff or the information inputter to re-verify and confirm the relevant information until accurate preoperative standard information is obtained.

[0082] In this embodiment, preoperative standard information refers to accurate, standardized, and complete preoperative information for children that meets the requirements of the pediatric preoperative assessment standard library after undergoing a series of processes including integrity verification, adaptive annotation, and secondary verification. This information provides a reliable basis for preoperative assessment of children, helps to formulate accurate surgical plans, and ensures the smooth progress of surgery.

[0083] like Figure 2 As shown, in order to obtain a child's preoperative basic information from the hospital management information system, medical staff mobile terminals, and parent self-service terminals through different methods, a multi-terminal interaction module is further proposed, including:

[0084] The hospital management information system's interactive submodule is used to call the structured data interface to directly read the basic fields of a single child's electronic medical record and obtain preoperative basic information from the hospital management information system.

[0085] The medical mobile terminal interaction submodule is used to input basic preoperative information of a single child from the medical mobile terminal based on a preset touch form;

[0086] The parent self-service terminal interaction submodule is used to input basic preoperative information of a single child from the parent self-service terminal based on the icon-based selection command entry of the preset graphic interaction interface.

[0087] In this embodiment, the structured data interface (SGA) is a standardized and well-defined data connection method used for data interaction between the hospital management information system and other systems or modules. It specifies the data format, transmission protocol, and access methods, enabling the hospital management information system to directly read relevant data from a single child's electronic medical record according to a specific structure and rules, ensuring data accuracy and consistency.

[0088] In this embodiment, the basic fields of the electronic medical record refer to the basic data units that record the child's preoperative basic information in the electronic medical record. These fields cover various basic information related to the child's medical care, such as name, age, gender, past medical history, allergy history, etc.

[0089] In this embodiment, the hospital management information system terminal is the core system terminal used by the hospital to manage and store patient medical information.

[0090] In this embodiment, a pre-designed touch-sensitive form is used specifically for information entry on medical mobile terminals. Medical staff enter basic preoperative information for individual children on this form via touch operation. The form will have different input boxes, drop-down menus, radio buttons, and other interactive elements set according to needs, making it convenient for medical staff to quickly and accurately input relevant information such as the child's current physical condition and various examination indicators.

[0091] In this embodiment, the medical mobile terminal typically refers to a mobile device, such as a tablet or smartphone, carried by medical staff during medical work. During the preoperative information entry process for children, medical staff use these devices to enter basic preoperative information for each child through a pre-set touch-screen form. The portability of this terminal allows medical staff to enter information promptly in different locations such as wards and examination rooms, improving the efficiency of information collection.

[0092] In this embodiment, a pre-defined graphical interface provides an icon-based selection command entry point: this is an information entry interaction method specifically designed for parent self-service terminals. On the pre-defined graphical interface, various information options are presented in an intuitive icon format. Parents can click on these icons to select the corresponding command and complete the entry of basic preoperative information for a single child. For example, there may be icons representing "allergy history," "past medical history," etc. Clicking on these icons expands the specific options for parents to choose from. This method reduces the difficulty of information entry for parents and improves the convenience of information entry.

[0093] In this embodiment, the parent self-service terminal is typically an electronic device used by parents, such as a mobile phone or tablet, to independently input relevant information about the child before surgery. Through a pre-defined graphical interface with icon-based command entry points, parents can easily input basic pre-operative information for their child, such as their daily health status and lifestyle habits. This supplements information not covered by the hospital management information system and medical staff mobile terminals, and is an important component of multi-terminal information input.

[0094] like Figure 3 As shown, in order to set mandatory field validation rules based on children's age groups, accurately detect missing preoperative basic information, and generate targeted completion prompts, a completeness validation module is further proposed, including:

[0095] The validation rule setting submodule is used to set the validation rules for the required fields of each preoperative core information dimension based on the child's age stratum.

[0096] The integrity verification submodule is used to detect whether there are any missing preoperative basic information entered by any terminal based on the mandatory field verification rules. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions based on all preoperative core information dimensions of children that do not have any missing information, until the complete preoperative information of a single child that has passed the verification is obtained.

[0097] In this embodiment, the mandatory field validation rules for each preoperative core information dimension are set based on the child's age stratum:

[0098] Because children of different ages have different physical conditions, common illnesses, and key information related to surgery, the system will stratify children according to their age. For example, children may be divided into different levels such as infancy (0-3 years old), preschool age (4-6 years old), school age (7-12 years old), and adolescence (13-18 years old).

[0099] For each core preoperative information dimension, such as basic health status, past medical history, and allergy history, corresponding mandatory field verification rules are set according to the child's age group. For example, for infants and toddlers, "birth weight" may be a mandatory field in the basic health status dimension because it is crucial for assessing the infant's development; while for adolescents, this field may not be mandatory, but "whether they have undergone sports training and related injury history" may be mandatory in certain surgical assessment scenarios. This ensures that the collected information is more targeted and closely related to the surgical needs of children of different age groups, providing accurate basic data for the development of surgical plans.

[0100] In this embodiment, based on the child's preoperative sub-basic information for all preoperative core information dimensions that are not missing, targeted completion prompts for the corresponding preoperative core information dimensions are generated until the complete preoperative information of a single child that has passed verification is obtained:

[0101] When the system detects missing information in a core preoperative information dimension, it generates targeted completion prompts using the child's preoperative sub-basic information from other core preoperative information dimensions that have already been fully entered. For example, suppose information is missing in the allergy history dimension, while the basic health status dimension is complete, recording the child's recent medication history. The system might generate a prompt based on this existing medication information: "Based on the child's recent medication history, there may be a risk of drug allergies. Please supplement the child's allergy history information, such as whether they are allergic to any recently taken medications." This process continuously generates targeted completion prompts, guiding the data entry user to supplement the missing information, and then re-verifies. This process is repeated until all core preoperative information dimensions are complete, resulting in a verified, complete preoperative information for a single child. This process ensures the completeness of preoperative information and avoids affecting the accuracy of the surgical plan due to the omission of crucial information.

[0102] like Figure 4 As shown, in order to determine the annotation granularity based on the type of input terminal, professional annotation of complete preoperative information is performed through steps such as symptom entity recognition and terminology graph retrieval. Furthermore, an adaptive annotation module is proposed, including:

[0103] The adaptive annotation granularity determination submodule is used to determine the adaptive annotation granularity based on the type of terminal used to input complete preoperative information.

[0104] The symptom entity recognition submodule is used to perform symptom entity recognition on the complete preoperative information of a single child to obtain all symptom entities of the corresponding child.

[0105] The terminology graph retrieval submodule is used to generate a pediatric terminology knowledge graph based on a pediatric terminology database, and to determine the matching standard terms for all non-professional input information of all symptom entities of a single child based on the annotation granularity and the pediatric terminology knowledge graph.

[0106] The matching information standard submodule is used to annotate all non-professional input information of a single child with matching standard terms and then to all non-professional input information in the corresponding child's complete preoperative information to obtain the corresponding child's preoperative annotation information.

[0107] In this embodiment, symptom entity recognition is performed on the complete preoperative information of a single child to obtain all symptom entities for the corresponding child:

[0108] This step aims to identify, using specific algorithms or techniques, the symptom-related components describing a child's health condition from the complete preoperative information obtained for a single child, and abstract them into "symptom entities." These symptom entities serve as the basic units for subsequent annotation and analysis.

[0109] For example, if the preoperative information includes a description such as "the child has been coughing recently and also said that his stomach hurts," the system can use natural language processing technology or predefined rules to identify "cough" and "stomach hurts" as two symptom entities, thereby obtaining all the symptom entities of the corresponding child, so as to further match and label them with professional terms.

[0110] In this embodiment, the pediatric terminology knowledge graph is a knowledge structure built upon a pediatric terminology database. It graphically displays the relationships between various terminology entities (such as specialist symptom terminology entities) within the pediatric medical field. This graph includes not only professional and non-professional descriptions of each specialist symptom terminology entity but also the correlation coefficients between all symptom entities. For example, in the graph, the specialist symptom terminology entity "fever" may have varying degrees of association with other symptom entities such as "cold" and "infection," with the correlation coefficient quantifying this relationship. Simultaneously, "fever" may have professional descriptions such as "pathological elevation of body temperature exceeding the normal range" as well as some common non-professional descriptions such as "the child has a fever." The knowledge graph provides an important reference framework for determining matching standard terms for symptom entities, helping the system to more accurately understand and process symptom descriptions in preoperative information about children.

[0111] In this embodiment, all non-professional information entered for the symptom entity refers to symptom-related content expressed in plain, non-medical language within a single child's complete preoperative information. Since the information may originate from non-professional input methods such as parent self-service terminals, it contains a large amount of everyday descriptions. For example, a parent might describe "my child has developed many small red bumps," where "many small red bumps" is non-professional information related to the symptom entity "rash." These non-professional expressions need to be converted into professional terminology through matching with a pediatric specialist terminology knowledge graph to improve the standardization and professionalism of the information.

[0112] In this embodiment, the matching standard terminology for non-professional input information refers to the corresponding professional medical terminology found for the symptom entity based on the pediatric terminology knowledge graph. The system determines the most suitable professional term based on the matching degree calculation between the non-professional input information and various descriptions in the knowledge graph. For example, for the aforementioned non-professional input information "many small red bumps on the body," by searching in the pediatric terminology knowledge graph and combining semantic similarity calculations, its matching standard term is determined to be "rash." Labeling non-professional input information as matching standard terms ensures that preoperative information conforms to medical professional standards, facilitating accurate understanding by medical staff and use in surgical assessments and other work.

[0113] To build a knowledge graph based on a pediatric specialty terminology database and retrieve matching standard terms through processes such as calculating matching degree and determining professional search depth, a terminology graph retrieval submodule is further proposed, including:

[0114] The knowledge graph building unit is used to determine all professional and non-professional descriptions of all symptom terminology entities based on the pediatric specialist terminology database. Based on the correlation coefficients between all symptom entities involved in the pediatric specialist terminology database and all professional and non-professional descriptions of all symptom terminology entities, a pediatric specialist terminology knowledge graph is generated.

[0115] The matching degree calculation unit is used to determine the matching degree between each symptom entity and each specialty symptom term entity based on all non-professional and professional input information of all symptom entities and all non-professional and professional description information in the pediatric specialty terminology knowledge graph.

[0116] The centralized positioning unit is used to perform centralized positioning in the pediatric specialty terminology knowledge graph based on the matching degree between each symptom entity and each specialty symptom term entity, and to determine the current centralized location range in the pediatric specialty terminology knowledge graph.

[0117] The professional retrieval depth determination unit is used to determine the professional retrieval depth based on the current concentration range and adaptive annotation granularity in the pediatric specialty terminology knowledge graph.

[0118] The standard term matching unit is used to retrieve matching standard terms for all non-professional input information of all symptom entities in the specialist symptom standard term conversion database based on professional retrieval depth.

[0119] In this embodiment, all professional and non-professional descriptive information of the specialist symptom terminology entity is included:

[0120] Professional descriptive information refers to the precise and standardized expression of specialized symptom terms within the field of pediatric medical medicine. These descriptions are based on medical expertise, industry standards, and clinical experience, and are therefore professional and accurate. For example, for the specialized symptom term "pneumonia," a professional description might be "inflammation of the terminal airways, alveoli, and pulmonary interstitium, which can be caused by pathogenic microorganisms, physical and chemical factors, immune damage, allergies, and medications," detailing the pathological location and causes of the disease.

[0121] Non-technical descriptive information: This refers to descriptions of specialized symptom terms in plain, everyday language, so that non-medical professionals (such as parents) can understand and express them. For "pneumonia," a non-technical description might be, "My child has lung inflammation, and keeps coughing and has a fever," describing the symptoms in a way that is closer to common understanding. The system collects and organizes these different types of descriptive information to match and label non-technical data with professional terminology when processing preoperative information.

[0122] In this embodiment, the correlation coefficient between symptom entities is a quantitative indicator used to measure the degree of association between different symptom entities in a pediatric terminology database. Different symptom entities may have various medical relationships, such as causal relationships and accompanying relationships, and the correlation coefficient reflects the strength of these relationships. For example, there may be a high correlation coefficient between the symptom entities "cough" and "cold" because colds are often accompanied by cough symptoms; while the correlation coefficient between "cough" and "fracture" is relatively low because there is less direct association between them. The correlation coefficient can be obtained by organizing experts in the field of pediatric medicine to score the degree of association between symptom entities based on their rich clinical experience and professional knowledge of pediatric diseases, or by reviewing a large amount of pediatric medical literature, including academic papers and clinical guidelines. By analyzing the relationship between symptoms and diseases in the literature, the correlation information between symptom entities can be extracted. For example, a reputable paper points out that in studies of childhood pneumonia, symptoms such as "fever," "cough," and "rapid breathing" have a high probability of occurring simultaneously, thus determining that these symptom entities have a high correlation coefficient.

[0123] In this embodiment, the current concentration area in the pediatric terminology knowledge graph represents a region in the knowledge graph that is closely related to the currently processed symptom entity. Further retrieval and analysis within this region helps to more accurately determine the matching standard terms for non-professional information, narrow the search scope, and improve matching efficiency and accuracy.

[0124] In this embodiment, the professional search depth is an important parameter when performing searches within the pediatric terminology knowledge graph. It determines the level of detail and scope of the search. It is determined based on the current concentration area within the pediatric terminology knowledge graph and the adaptive annotation granularity. An appropriate professional search depth ensures search accuracy while avoiding resource waste and efficiency reduction caused by excessive searching, ensuring that the system can quickly and accurately find standard terms that match the non-professional information entered for symptom entities from a large amount of terminology information.

[0125] In this embodiment, based on the professional retrieval depth, matching standard terms for all non-professional input information of all symptom entities are retrieved from the specialized symptom standard terminology conversion database:

[0126] The Specialty Symptom Standard Terminology Conversion Database is a database that stores the correspondence between a large amount of non-professional input information of symptom entities and professional standard terminology. The system performs retrieval operations in this database based on a determined professional search depth.

[0127] Based on the determined professional search depth, a search is conducted in the specialized symptom standard terminology conversion database. The professional search depth determines the scope and level of detail of the search. For example, a professional search depth of 0.38 means that the database will search for standard terms matching the non-professional entry information for each symptom entity with a specific range and precision. If the search is for the non-professional entry information "child's stomachache," the system will find the most matching standard term "abdominal pain" in the conversion database based on the scope defined by the professional search depth.

[0128] In this embodiment, the Specialty Symptom Standard Terminology Conversion Database is a database specifically designed to store and manage the correspondence between non-professional input information of symptom entities and professional standard terminology within the field of pediatric specialty. It is a key component for converting non-professional input information into professional terminology. The database contains a wealth of entries, each recording one or more non-professional input information entries and their corresponding professional standard terminology. For example, for the symptom "headache," the database might record non-professional input information such as "headache" or "severe headache," corresponding to the professional standard term "headache."

[0129] To determine the matching degree between symptom entities and specialized symptom terminology entities by calculating the semantic similarity between non-professional and professional descriptive information, a matching degree calculation unit is further proposed, comprising:

[0130] Based on the semantic similarity between each non-professional entry information of each symptom entity and each non-professional description information of each specialty symptom term entity, and the semantic similarity between each professional entry information of each symptom entity and each professional description information of each specialty symptom term entity;

[0131] Based on the semantic similarity between all non-professional input information of each symptom entity and each non-professional description information of each specialist symptom term entity, and the semantic similarity between all professional input information of the corresponding symptom entity and each professional description information of the corresponding specialist symptom term entity, the matching degree between each symptom entity and the corresponding specialist symptom term entity is calculated.

[0132] In this embodiment, the semantic similarity between each non-professional entry information of each symptom entity and each non-professional description information of each specialist symptom term entity is as follows:

[0133] When processing children's preoperative information, the system compares each non-professional entry of the identified symptom entity with each non-professional description of each specialty symptom term entity in the pediatric specialty terminology knowledge graph to assess their semantic similarity.

[0134] For example, the symptom entity "stomach ache" might be entered as "The baby says his stomach hurts" in a non-professional context, while the specialized symptom term entity "abdominal pain" might be entered as "The child says his stomach is uncomfortable." Natural language processing techniques, such as word vector models (like Word2Vec and GloVe) or deep learning-based semantic understanding models (like BERT), are used to calculate the semantic similarity between these two sentences. These techniques consider factors such as word meaning and contextual relationships, providing a numerical value to represent their similarity; a higher value indicates greater semantic similarity.

[0135] In this embodiment, the semantic similarity between each specialty entry for each symptom entity and each specialty description for each specialty symptom term entity is as follows:

[0136] Similar to the processing of non-professional information, for professionally entered information that may exist in symptom entities (such as more professional symptom descriptions obtained from the hospital management information system), the system will also perform semantic similarity calculations between them and the professional description information of the specialized symptom terminology entities.

[0137] Suppose the symptom entity "fever" is entered as "body temperature 38.5℃, persistent fever for 2 days," while the professional description of the specialized symptom term "fever" is "a body temperature exceeding the normal range and maintained for a certain period of time." Using professional medical terminology understanding and analysis tools, combined with a medical knowledge system, we can assess the semantic similarity between these two descriptions.

[0138] In this embodiment, based on the semantic similarity between all non-professional input information of each symptom entity and each non-professional description information of each specialist symptom term entity, and the semantic similarity between all professional input information of the corresponding symptom entity and each professional description information of the corresponding specialist symptom term entity, the matching degree between each symptom entity and the corresponding specialist symptom term entity is calculated:

[0139] To comprehensively and accurately determine the degree of matching between symptom entities and specialized symptom terminology entities, the system will comprehensively consider semantic similarity at both the non-professional and professional levels. In the specific calculation, it may perform weighted summation and other operations based on different weights on the semantic similarity between all non-professional input information and non-professional description information, as well as the semantic similarity between professional input information and professional description information.

[0140] For example, if we assume that non-professional semantic similarity accounts for 40% of the weight in the matching score, and professional semantic similarity accounts for 60%, then for a certain symptom entity and its corresponding specialist symptom term entity, assuming that the average semantic similarity between all non-professional entries and non-professional descriptions is 0.7, and the average semantic similarity between all professional entries and professional descriptions is 0.8, then the weighted matching score can be calculated as: 0.7 × 40% + 0.8 × 60% = 0.76.

[0141] To determine the current focal location range in the pediatric specialty terminology knowledge graph through operations such as screening highly matching entities and locating temporary associated subnetworks, a focal location unit is further proposed, including:

[0142] The high-match entity filtering subunit is used to filter out all specialist symptom terminology entities in the pediatric specialist terminology knowledge graph whose match degree with each symptom entity exceeds a preset match degree threshold as all high-match entities of each symptom entity.

[0143] Temporary subnetwork localization subunits are used to identify temporary association subnetworks in the pediatric terminology knowledge graph based on the direct association paths between all highly matched entities of all symptom entities.

[0144] The first centralized location subunit is used to treat the temporary associated subnetwork as the current centralized location range in the children's specialty terminology knowledge graph when the average association coefficient of the temporary associated subnetwork is greater than the first association coefficient threshold and the node clustering coefficient is greater than the second association threshold.

[0145] The second centralized location subunit is used to perform a stepwise increase in the matching degree when the average correlation coefficient of the temporary associated subnetwork is not greater than the first correlation coefficient threshold or the node clustering coefficient is not greater than the second correlation coefficient threshold. Based on the latest matching value, all high-matching entities of each symptom entity are re-filtered until the average correlation coefficient of the temporary associated subnetwork determined based on all high-matching entities of all the latest filtered symptom entities is greater than the first correlation coefficient threshold and the node clustering coefficient is greater than the second correlation coefficient threshold. Then, the newly determined temporary associated subnetwork is regarded as the current centralized location range in the pediatric specialty terminology knowledge graph.

[0146] In this embodiment, a preset matching threshold is used: this is a pre-defined numerical standard used to determine whether the matching degree between a symptom entity and a specialist symptom term entity is sufficiently high. For example, if the preset matching threshold is set to 0.6, when the matching degree between a symptom entity and a specialist symptom term entity is greater than 0.6, they are considered to have a high degree of matching, and the specialist symptom term entity may be regarded as a high-matching entity. Setting this threshold helps to filter out specialist symptom term entities that are strongly associated with the symptom entity, thereby narrowing the scope of analysis and improving the accuracy and efficiency of determining the matching standard terms.

[0147] In this embodiment, the direct association path between highly matched entities refers to the direct connection between these highly matched entities. For example, in a knowledge graph, "cough" and "cold" might be highly matched entities for a certain symptom entity, and there is a direct association between them such as "colds often cause coughs." The path formed by this association is the direct association path between highly matched entities. Understanding these direct association paths helps to identify a closely related region in the knowledge graph, i.e., a temporary association subnetwork, so as to more accurately analyze the relationship between symptom entities and specialized symptom terminology entities.

[0148] In this embodiment, a temporary associated subnetwork is identified in the pediatric terminology knowledge graph based on the direct association paths between all highly matched entities of all symptom entities. After calculating the matching degree between each symptom entity and the specialized symptom terminology entity, and filtering out the highly matched entities of all symptom entities, a local network structure is delineated in the entire pediatric terminology knowledge graph according to the direct association paths between these highly matched entities. This structure is the temporary associated subnetwork. For example, assuming there are multiple symptom entities, each of which has several highly matched entities, and these highly matched entities are interconnected through direct association paths, the region formed by these interconnected highly matched entities and their association paths is identified from the entire knowledge graph, thus forming the temporary associated subnetwork.

[0149] In this embodiment, the average association coefficient and node clustering coefficient of the temporary associated subnetwork are:

[0150] Average correlation coefficient: This is an average value that measures the strength of the association between symptom entities in a temporary association subnetwork. Since the correlation coefficient represents the strength of the relationship between symptom entities, the average correlation coefficient is the sum of the correlation coefficients between all symptom entities in the temporary association subnetwork, divided by the total number of associations. A higher average correlation coefficient means that the symptom entities within that subnetwork generally have strong associations. For example, if there are three symptom entities A, B, and C in a temporary association subnetwork, and the correlation coefficient between A and B is 0.8, the correlation coefficient between A and C is 0.7, and the correlation coefficient between B and C is 0.9, then the average correlation coefficient is (0.8 + 0.7 + 0.9) ÷ 3 = 0.8.

[0151] Node clustering coefficient: This describes the degree of clustering of nodes (i.e., symptom entities) in a temporary associated subnetwork, reflecting the tightness of connections between a node's neighboring nodes. Specific calculation methods include:

[0152] Determine the neighboring nodes of a single node, calculate the actual number of edges between all neighboring nodes of a single node, calculate the maximum possible number of edges between all neighboring nodes of a single node (this value represents the number of edges if all neighboring nodes are interconnected, i.e., the maximum number of edges), and take the average of the ratios of the actual number of edges between all neighboring nodes of each node to the maximum possible number of edges between all neighboring nodes of the corresponding node as the clustering coefficient of the single node.

[0153] The average of the clustering coefficients of all nodes in the temporary associated subnetwork is taken as the node clustering coefficient.

[0154] In this embodiment, the first correlation coefficient threshold is a pre-set limit value for the average correlation coefficient. For example, the first correlation coefficient threshold is set to 0.7. If the average correlation coefficient of the temporary correlated subnetwork is 0.8, which is greater than the threshold, then the subnetwork meets the preliminary requirements for the current concentration location range in terms of the average correlation coefficient.

[0155] In this embodiment, the second correlation coefficient threshold is a pre-set limit value for the node clustering coefficient. For example, if the second correlation coefficient threshold is set to 0.6, and the node clustering coefficient of the temporary associated subnetwork is 0.7, which is greater than the threshold, then from the perspective of the node clustering coefficient, the subnetwork also meets one of the conditions for being the current concentrated location range.

[0156] In this embodiment, the matching degree is increased in a stepwise manner, and all high-matching entities for each symptom entity are re-filtered based on the latest matching value. When the average correlation coefficient of the temporary associated subnetwork is not greater than the first correlation coefficient threshold or the node clustering coefficient is not greater than the second correlation coefficient threshold, it indicates that the temporary associated subnetwork formed by the currently selected high-matching entities may not be the core region most closely related to the symptom entity. At this time, the system will increase the matching degree between the symptom entity and the specialized symptom terminology entity in a stepwise manner, that is, increase the matching degree by a certain step size (e.g., 0.05) each time. Then, based on the latest matching value after the increase, the matching degree of each symptom entity is compared with all specialized symptom terminology entities again, and all specialized symptom terminology entities with matching degrees exceeding the preset matching degree threshold are selected as high-matching entities for each symptom entity. In this way, the filtering conditions are continuously adjusted in order to find a temporary associated subnetwork with an average correlation coefficient greater than the first correlation coefficient threshold and a node clustering coefficient greater than the second correlation coefficient threshold, thereby accurately determining the current concentration range in the pediatric specialized terminology knowledge graph and laying the foundation for subsequent accurate retrieval and matching of standard terms. For example, if the initial matching degree is 0.6 and the preset matching degree threshold is 0.6, the temporary associated sub-network formed by the selected high matching degree entities does not meet the requirements. The matching degree is increased by 0.05 to 0.65, and high matching degree entities are re-selected. If the new temporary associated sub-network meets the conditions, the subsequent operations are continued based on this.

[0157] To determine the professional retrieval depth based on location vectors and adaptive annotation granularity through processes such as selecting core term entities and expanding multi-level term entity groups, a professional retrieval depth determination unit is further proposed, including:

[0158] The core term entity selection sub-unit is used to determine the comprehensive matching degree between each specialty symptom term entity and all symptom entities in the current centralized location range, and the specialty symptom term entity with the largest comprehensive matching degree in the current centralized location range is regarded as the core term entity.

[0159] Multi-level term entity extension subunits are used to identify multiple levels of extended entity groups from the current centralized location range, centered on the core term entity and in descending order of the correlation coefficient.

[0160] The location vector generation subunit is used to generate the location vector of each level of the extended entity group based on the distance between all edge nodes of each level of the extended entity group and all edge nodes of the children's specialty terminology knowledge graph.

[0161] The retrieval depth determination subunit is used to determine the specialized retrieval depth based on the location vectors and adaptive annotation granularity of the extended entity group at all levels.

[0162] In this embodiment, the overall matching degree between each specialty symptom term entity and all symptom entities within the current centralized location range is determined:

[0163] Different weights are assigned based on the actual situation for calculation. Assuming the non-professional semantic similarity weight is 0.4 and the professional semantic similarity weight is 0.6, for a specific specialty symptom terminology entity, its average non-professional semantic similarity with all symptom entities is 0.7, and its average professional semantic similarity is 0.8. Therefore, the overall matching degree between this specialty symptom terminology entity and all symptom entities is 0.7 × 0.4 + 0.8 × 0.6 = 0.76. By calculating the overall matching degree, we can identify the specialty symptom terminology entity that is most closely associated with all symptom entities within the current concentrated location range; this is the core terminology entity, providing a core reference point for subsequent operations.

[0164] In this embodiment, taking the core term entity as the center, multiple levels of extended entity groups are sequentially identified from the current centralized location range according to the principle of descending order of correlation coefficient:

[0165] After identifying the core term entity, it serves as the center, and related entities are expanded based on the correlation coefficients between symptom entities. The correlation coefficient reflects the degree of association between different symptom entities. Following a descending order of correlation coefficients, this means starting with the entity with the highest correlation coefficient to the core term entity, a certain number of entities are selected sequentially to form different levels of expanded entity groups. For example, if the core term entity is "pneumonia," the entities with the highest correlation coefficients to it might be "cough" and "fever," which are designated as the first-level expanded entity group. Next, entities with slightly lower correlation coefficients to the core term entity, such as "difficulty breathing" and "sputum production," are selected to form the second-level expanded entity group, and so on.

[0166] In this embodiment, a location vector for each level of the extended entity group is generated based on the distances between all edge nodes of each level's extended entity group and all edge nodes of the children's specialty terminology knowledge graph:

[0167] Identify all edge nodes of each level of extended entity group. These edge nodes represent the boundary positions of that extended entity group in the knowledge graph. Calculate the distances between these edge nodes and all edge nodes of the pediatric terminology knowledge graph. These distances can be measured in various ways, such as by calculating the path length between nodes within the knowledge graph structure. Organizing and digitizing this distance information generates the location vector for each level of extended entity group. For example, for a first-level extended entity group, the distances between its edge nodes and all edge nodes of the knowledge graph, after calculation and organization, might be represented as a vector form like [0.2, 0.3, 0.5, ...].

[0168] To determine the overall professional search depth by calculating the first sub-specialty search depth corresponding to different levels of extended entity groups, determining the second sub-specialty search depth based on rules, and then further determining the professional search depth, a search depth determination sub-unit is proposed, including:

[0169] The first retrieval depth determination end is used to determine the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth. Based on the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth and the positioning vector of the extended entity group at each level, the first sub-professional retrieval depth corresponding to the extended entity group at each level is calculated.

[0170] The second retrieval depth determination end is used to determine the second sub-specialty retrieval depth based on the adaptive annotation granularity and the annotation granularity-specialty retrieval depth mapping rule;

[0171] The third search depth determination end is used to determine the professional search depth based on the search depths of all first sub-professions and second sub-professions.

[0172] In this embodiment, the mapping coefficient vector between the location vector of each level of extended entity group and the professional retrieval depth is determined:

[0173] To establish the connection between the two, a mapping coefficient vector needs to be determined. This mapping coefficient vector is derived through analysis and calculation, and each element represents the proportional relationship or conversion factor between the corresponding dimension of the positioning vector and the professional retrieval depth. For example, assuming the positioning vector is a three-dimensional vector [x, y, z], after a series of data analyses and model calculations, the mapping coefficient vector is determined to be [a, b, c]. Here, 'a' represents the degree of correlation between the x dimension of the positioning vector and the professional retrieval depth, and 'b' and 'c' are similarly determined.

[0174] In this embodiment, based on the mapping coefficient vector between the location vector of each level of extended entity group and the professional retrieval depth, and the location vector of each level of extended entity group, the first sub-professional retrieval depth corresponding to each level of extended entity group is calculated:

[0175] Once the mapping coefficient vector and the location vector of the extended entity group are determined, the first sub-specialty retrieval depth can be calculated. The specific calculation method involves multiplying each dimension of the location vector by the corresponding dimension of the mapping coefficient vector, and then summing the results (or performing other mathematical operations according to specific calculation rules). For example, for an extended entity group at a certain level, with location vectors [x1, x2, x3] and mapping coefficient vectors [a1, a2, a3], the first sub-specialty retrieval depth D1 = a1x1 + a2x2 + a3x3. This first sub-specialty retrieval depth reflects the retrieval depth determined based on the position of the extended entity group in the knowledge graph. Different levels of extended entity groups will calculate different first sub-specialty retrieval depths, which provide an important reference value for finally determining the specialty retrieval depth from a positional perspective.

[0176] In this embodiment, the annotation granularity-specialty retrieval depth mapping rule is a pre-defined rule describing the correspondence between annotation granularity and specialty retrieval depth. Specialty retrieval depth refers to the level of detail required to retrieve matching standard terms within the knowledge graph. This mapping rule specifies the appropriate specialty retrieval depth for different annotation granularities. For example, when the annotation granularity is coarse, the specialty retrieval depth is correspondingly shallow, potentially only retrieving within a smaller area of ​​the knowledge graph; conversely, when the annotation granularity becomes finer, the specialty retrieval depth should be increased, requiring retrieval within a more detailed and broader range of the knowledge graph. This rule provides a standardized reference for determining specialty retrieval depth based on adaptive annotation granularity.

[0177] In this embodiment, the second sub-specialty retrieval depth is determined based on the adaptive annotation granularity and the annotation granularity-specialty retrieval depth mapping rule:

[0178] Adaptive annotation granularity is the level of detail for annotating non-specialist information, determined by the type of terminal used to input preoperative complete information. Based on this adaptive annotation granularity, and referring to the annotation granularity-specialty retrieval depth mapping rule, the second sub-specialty retrieval depth can be determined. For example, if the adaptive annotation granularity indicates that more detailed annotation is needed, the corresponding specialty retrieval depth value is found as the second sub-specialty retrieval depth according to the mapping rule. Suppose the mapping rule stipulates that when the adaptive annotation granularity corresponds to a finer granularity, the specialty retrieval depth should be set to a specific value; then this is the second sub-specialty retrieval depth.

[0179] In this embodiment, the professional search depth is determined based on the search depth of all first sub-professions and second sub-professions:

[0180] To obtain a final, accurate, and reasonable professional search depth, it is necessary to comprehensively consider the search depths of both sub-specialties. For example, the specific determination method might be through an algorithm that takes the maximum or minimum value.

[0181] To generate assessment items based on the preoperative annotation information and associated assessment standard library, identify conflicting information entered across terminals, determine the verification path to trigger secondary verification prompts, a secondary verification module is further proposed, including:

[0182] The assessment item generation submodule is used to automatically generate structured assessment items by associating with the preoperative assessment standard library for children based on preoperative annotation information;

[0183] The conflict identification submodule is used to identify cross-terminal input conflict information in the preoperative annotation information and associate conflicting terminals;

[0184] The verification path determination submodule is used to determine the verification path based on structured evaluation items, cross-terminal input conflict information, and conflicting terminals.

[0185] The verification prompt submodule is used to trigger secondary verification prompts based on the verification path until the preoperative standard information is obtained.

[0186] In this embodiment, the conflicting terminals are associated: during the preoperative information entry process for children, when the system identifies cross-terminal input conflict information in the preoperative annotation information, it will associate this conflicting information with the specific terminal that caused the conflict. This is called "associating conflicting terminals".

[0187] In this embodiment, the verification path is determined based on structured assessment items, conflicting information entered across terminals, and the conflicting terminals. Determining the verification path based on these three factors involves planning a reasonable information verification process according to the specific circumstances of the conflict and the assessment criteria. For example, if the structured assessment items indicate that the child's past medical history is crucial for surgical decisions, but there is a conflict between the information entered in the hospital management information system and the parent's self-service terminal, then the verification path might first contact the parents to confirm the basis and accuracy of the information entered in the parent's self-service terminal. Simultaneously, medical staff would re-verify the information source in the hospital management information system, such as by reviewing relevant medical records. This verification path is determined based on the importance of the conflicting information (reflected by the structured assessment items) and the characteristics of the conflicting terminals, aiming to efficiently resolve information conflicts through a structured process and ensure accurate preoperative information.

[0188] In this embodiment, secondary verification prompts are triggered based on the verification path until preoperative standard information is obtained. Once the verification path is determined, the system will trigger secondary verification prompts according to this path. Secondary verification prompts are notifications sent to relevant personnel (such as medical staff, parents, etc.) to re-verify information. For example, following the verification path described above, the system may send a prompt to parents informing them that the child's past medical history information they entered conflicts with the hospital system records, requesting them to confirm and provide accurate information again; simultaneously, it will prompt medical staff to verify relevant information in the hospital's management information system. After relevant personnel verify and correct the information according to the prompts, the system will conduct another assessment and check. If there are still discrepancies with the preoperative standard information requirements, secondary verification prompts will continue to be triggered according to the verification path, repeating this process until preoperative standard information that meets the requirements of the pediatric preoperative assessment standard database is obtained. By continuously triggering secondary verification prompts based on the verification path, the accuracy and standardization of preoperative information can be ensured, providing a reliable guarantee for the smooth progress of pediatric surgery.

[0189] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-terminal adaptive annotation system for preoperative information entry in children, characterized in that, include: A multi-terminal interaction module is used to receive preoperative basic information of a single child from multiple pre-adapted terminal sources; The integrity verification module is used to detect whether there are any missing preoperative basic information entered by any terminal based on the verification rules of the required fields of all preoperative core information dimensions. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions until the complete preoperative information of a single child that has passed the verification is obtained. The adaptive annotation module is used to determine the adaptive annotation granularity based on the type of input terminal for complete preoperative information, and to dynamically match and annotate non-professional input information in the complete preoperative information of a single child based on the pediatric terminology database and the adaptive annotation granularity, so as to obtain the preoperative annotation information of a single child. The secondary verification module is used to associate the preoperative assessment standard library for children with the preoperative annotation information and generate structured assessment items. It identifies cross-terminal input conflict information in the preoperative annotation information and triggers secondary verification prompts based on the structured assessment items and cross-terminal input conflict information until the preoperative standard information is obtained. The adaptive annotation module includes: The adaptive annotation granularity determination submodule is used to determine the adaptive annotation granularity based on the type of input terminal for complete preoperative information. The symptom entity recognition submodule is used to perform symptom entity recognition on the complete preoperative information of a single child to obtain all symptom entities of the corresponding child. The terminology graph retrieval submodule is used to generate a pediatric terminology knowledge graph based on a pediatric terminology database, and to determine the matching standard terms for all non-professional input information of all symptom entities of a single child based on the annotation granularity and the pediatric terminology knowledge graph. The matching information standard submodule is used to annotate all the non-professional information entered by a single child with the matching standard terms and then to all the non-professional information entered by the corresponding child in the complete preoperative information, so as to obtain the preoperative annotation information of the corresponding child. The terminology map retrieval submodule includes: The knowledge graph building unit is used to determine all professional and non-professional descriptions of all symptom terminology entities based on the pediatric specialist terminology database. Based on the correlation coefficients between all symptom entities involved in the pediatric specialist terminology database and all professional and non-professional descriptions of all symptom terminology entities, a pediatric specialist terminology knowledge graph is generated. The matching degree calculation unit is used to determine the matching degree between each symptom entity and each specialty symptom term entity based on all non-professional and professional input information of all symptom entities and all non-professional and professional description information in the pediatric specialty terminology knowledge graph. The centralized positioning unit is used to perform centralized positioning in the pediatric specialty terminology knowledge graph based on the matching degree between each symptom entity and each specialty symptom term entity, and to determine the current centralized location range in the pediatric specialty terminology knowledge graph. The professional retrieval depth determination unit is used to determine the professional retrieval depth based on the current concentration range and adaptive annotation granularity in the pediatric specialty terminology knowledge graph. The standard term matching unit is used to retrieve matching standard terms for all non-professional input information of all symptom entities in the specialist symptom standard term conversion database based on professional retrieval depth.

2. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, Multi-terminal interaction module, including: The hospital management information system's interactive submodule is used to call the structured data interface to directly read the basic fields of a single child's electronic medical record and obtain preoperative basic information from the hospital management information system. The medical mobile terminal interaction submodule is used to input basic preoperative information of a single child from the medical mobile terminal based on a preset touch form; The parent self-service terminal interaction submodule is used to input basic preoperative information of a single child from the parent self-service terminal based on the icon-based selection command entry of the preset graphic interaction interface.

3. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, The integrity verification module includes: The validation rule setting submodule is used to set the validation rules for the required fields of each preoperative core information dimension based on the child's age stratum. The integrity verification submodule is used to detect whether there are any missing preoperative basic information entered by any terminal based on the mandatory field verification rules. If so, it generates targeted completion prompts for the corresponding preoperative core information dimensions based on all preoperative core information dimensions of children that do not have any missing information, until the complete preoperative information of a single child that has passed the verification is obtained.

4. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, The matching degree calculation unit includes: Based on the semantic similarity between each non-professional entry information of each symptom entity and each non-professional description information of each specialty symptom term entity, and the semantic similarity between each professional entry information of each symptom entity and each professional description information of each specialty symptom term entity; Based on the semantic similarity between all non-professional input information of each symptom entity and each non-professional description information of each specialist symptom term entity, and the semantic similarity between all professional input information of the corresponding symptom entity and each professional description information of the corresponding specialist symptom term entity, the matching degree between each symptom entity and the corresponding specialist symptom term entity is calculated.

5. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, Centralized positioning unit, including: The high-match entity filtering subunit is used to filter out all specialist symptom terminology entities in the pediatric specialist terminology knowledge graph whose match degree with each symptom entity exceeds a preset match degree threshold as all high-match entities of each symptom entity. Temporary subnetwork localization subunits are used to identify temporary association subnetworks in the pediatric terminology knowledge graph based on the direct association paths between all highly matched entities of all symptom entities. The first centralized location subunit is used to treat the temporary associated subnetwork as the current centralized location range in the children's specialty terminology knowledge graph when the average association coefficient of the temporary associated subnetwork is greater than the first association coefficient threshold and the node clustering coefficient is greater than the second association threshold. The second centralized location subunit is used to perform a stepwise increase in the matching degree when the average correlation coefficient of the temporary associated subnetwork is not greater than the first correlation coefficient threshold or the node clustering coefficient is not greater than the second correlation coefficient threshold. Based on the latest matching value, all high-matching entities of each symptom entity are re-filtered until the average correlation coefficient of the temporary associated subnetwork determined based on all high-matching entities of all the latest filtered symptom entities is greater than the first correlation coefficient threshold and the node clustering coefficient is greater than the second correlation coefficient threshold. Then, the newly determined temporary associated subnetwork is regarded as the current centralized location range in the pediatric specialty terminology knowledge graph.

6. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, The specialized search depth determination unit includes: The core term entity selection sub-unit is used to determine the comprehensive matching degree between each specialty symptom term entity and all symptom entities in the current centralized location range, and the specialty symptom term entity with the largest comprehensive matching degree in the current centralized location range is regarded as the core term entity. Multi-level term entity extension subunits are used to identify multiple levels of extended entity groups from the current centralized location range, centered on the core term entity and in descending order of the correlation coefficient. The location vector generation subunit is used to generate the location vector of each level of the extended entity group based on the distance between all edge nodes of each level of the extended entity group and all edge nodes of the children's specialty terminology knowledge graph. The retrieval depth determination subunit is used to determine the specialized retrieval depth based on the location vectors and adaptive annotation granularity of the extended entity group at all levels.

7. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 6, characterized in that, The retrieval depth determines the sub-unit, including: The first retrieval depth determination end is used to determine the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth. Based on the mapping coefficient vector between the positioning vector of the extended entity group at each level and the professional retrieval depth and the positioning vector of the extended entity group at each level, the first sub-professional retrieval depth corresponding to the extended entity group at each level is calculated. The second retrieval depth determination end is used to determine the second sub-specialty retrieval depth based on the adaptive annotation granularity and the annotation granularity-specialty retrieval depth mapping rule; The third search depth determination end is used to determine the professional search depth based on the search depths of all first sub-professions and second sub-professions.

8. The multi-terminal adaptive annotation system for preoperative information entry in children according to claim 1, characterized in that, The secondary verification module includes: The assessment item generation submodule is used to automatically generate structured assessment items by associating with the preoperative assessment standard library for children based on preoperative annotation information; The conflict identification submodule is used to identify cross-terminal input conflict information in the preoperative annotation information and associate conflicting terminals; The verification path determination submodule is used to determine the verification path based on structured evaluation items, cross-terminal input conflict information, and conflicting terminals. The verification prompt submodule is used to trigger secondary verification prompts based on the verification path until the preoperative standard information is obtained.

Images