A resident rotation intelligent management interaction method and system

Abstract

The application provides a resident rotation intelligent management interaction method and system, relates to the medical education information technology field, and the application realizes full-process intelligent rotation management through the skill dependence graph atlas management, intelligent rotation plan and dynamic feedback control of the main interface integration: first, a graph atlas expressing the front and rear dependencies of clinical skills is constructed, and a visual skill tree is automatically generated for each resident; then, the multi-dimensional data, skill state, rotation period and teaching resource constraints are fused, a compliant rotation scheme is intelligently generated and displayed, interactive adjustment and real-time verification of hard constraints are supported; feedback is continuously collected during execution, a comprehensive risk index is dynamically calculated, and the scheme is timely optimized, thereby forming a closed-loop management system covering plan development, execution monitoring and optimization iteration, and significantly improving the scientific nature, individualization level and operation efficiency of rotation management.

Description

Technical Field

[0001] This application relates to the field of medical education informatization technology, and in particular to an intelligent management and interaction method and system for resident physician rotation. Background Technology

[0002] Currently, large hospitals generally implement a standardized residency training system, requiring newly hired doctors to rotate through different clinical departments. However, the traditional rotation management model mainly relies on manual scheduling, paper records, and experience-based judgment, which has many problems: the rotation arrangement lacks scientific basis, is mostly based on fixed cycles, and does not fully consider key factors such as the individual ability, proficiency in operational skills, and workload of resident physicians. The rotation plan needs to be adjusted to adapt to the dynamic growth status and individualized training needs of resident physicians. Summary of the Invention

[0003] This application provides an intelligent management and interactive method and system for resident physician rotations to solve one or more technical problems existing in the prior art, and at least provides a beneficial option or creates conditions that can realize a personalized, dynamically adjustable, and teaching-compliant intelligent rotation management closed loop based on skill dependence and multi-source real-time feedback data.

[0004] On the one hand, this application provides an intelligent management and interaction method for resident physician rotation, including the following steps: The main interface displays the intelligent management interaction interface for resident physician rotations; wherein, the main interface includes a skill dependency graph management control, an intelligent rotation plan control, and a dynamic feedback control; In response to the trigger command of the skill dependency graph management control, the skill dependency graph management sub-interface is displayed, a skill dependency graph representing the logical dependencies between clinical operation skills is constructed and maintained, and a skill tree of the resident physician is generated and displayed based on the skill dependency graph. The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependency relationship between preceding and subsequent skills. In response to the trigger command of the intelligent rotation plan control, the intelligent rotation plan sub-interface is displayed, multi-dimensional input data of resident physicians is collected, and the rotation plan is generated and visualized by combining the skill tree, rotation cycle and teaching resource constraints, and interactive adjustment and hard constraint verification are supported; In response to the trigger command of the dynamic feedback control, a dynamic feedback sub-interface is displayed to collect feedback data in real time during the execution of the rotation scheme, calculate the comprehensive risk index, adjust the rotation scheme, and achieve closed-loop management.

[0005] Furthermore, the skill dependency graph management sub-interface includes a skill node addition control, a dependency configuration control, and a skill tree display area; The skill dependency graph management sub-interface displays the skill dependency graph management sub-interface, constructs and maintains a skill dependency graph representing the logical dependencies between clinical operation skills, and generates and displays the resident physician's skill tree based on the skill dependency graph, including the following steps: In response to the trigger command for adding a control to the skill node, the skill node editing window is displayed, and the basic information of the new skill input by the user is received. The basic information includes the skill name, the professional base, the required study hours and the qualification standards. The new skill is then added as an independent node to the skill dependency graph. In response to the trigger command of the dependency configuration control, the dependency configuration window is displayed. Directed dependency edges are established between two skill nodes by dragging or selecting. During the establishment process, it is checked whether a loop is formed. If a loop is detected, saving is prohibited and the user is prompted to correct it, so as to ensure that the skill dependency graph always maintains a directed acyclic structure. Based on the skill dependency graph, a skill tree is generated for each resident physician, which includes automatically matching the corresponding prerequisite skills and subsequent skill paths according to the skill items that the resident physician has completed. The skill tree display area shows the skill tree of the selected resident physician.

[0006] Furthermore, generating a skill tree for each resident physician based on the skill dependency graph includes the following steps: Obtain the set of skills that the current resident physician has completed; In the skill dependency graph, the nodes corresponding to the completed skill items are marked as mastered. Starting from all mastered nodes, traverse the skill dependency graph, identify all locked skill nodes whose prerequisite skills have been marked as mastered, and add the locked skill nodes to the unlockable skill set; Add the skill nodes from the unlockable skill set to the resident physician's skill tree and update the mastery status; Repeat the traversal and identification steps until the set of unlockable skills is empty, thereby generating a complete skill tree.

[0007] Furthermore, the skill dependency graph management sub-interface also includes version management controls; In response to the trigger command of the version management control, the version management window is displayed, which supports saving the current skill dependency graph version, publishing to the production environment, rolling back to the previous version, and branch management and independent maintenance of the skill dependency graph according to different professional bases.

[0008] Furthermore, the intelligent rotation plan sub-interface includes a data import control, an operation video upload control, a rotation plan generation control, and a plan adjustment control; The intelligent rotation plan sub-interface collects multi-dimensional input data from resident physicians, combines it with the skill tree, rotation cycle, and teaching resource constraints, generates and visualizes the rotation plan, and supports interactive adjustments and hard constraint verification, including the following steps: In response to the trigger command of the data import control, a data import window is displayed, and an electronic medical record system, simulation training platform or mobile scoring interface is configured and connected to obtain the resident physician's theoretical learning status, teaching score and personal basic information. In response to the trigger command of the operation video upload control, a video acquisition window is displayed to receive clinical operation videos uploaded by resident physicians or teaching staff, perform posture estimation and instrument trajectory recognition, and generate structured skill scores. In response to the trigger command of the rotation scheme generation control, based on the skill tree, the skill score, the upper limit of the rotation cycle, the department capacity limit and the status of teaching resources, the rule-based scheduling engine is invoked to generate a rotation scheme that meets the hard constraints of the prerequisite skill achievement, and the scheme is visualized in the form of a Gantt chart. In response to the trigger command of the scheme adjustment control, the user is allowed to drag and adjust the rotation period, and the scheme after adjustment is checked in real time to see if it violates hard constraints. When there is a conflict, a prompt message will pop up to ensure the compliance and executability of the rotation plan.

[0009] Furthermore, the process of invoking the rule-based scheduling engine to generate a round-robin scheme includes the following steps: Based on the skill tree and the skill scores, the set of rotation departments that each resident physician can currently enter is determined. A department is only included in the set of rotation departments that can be entered when the skill scores corresponding to all the prerequisite skills required by a department reach a preset threshold. Based on the preset rotation cycle and the maximum number of people that can be received in each department, the rotation time slots for resident physicians are allocated according to the first-come, first-served or preset rotation priority order. Synchronously match teaching resources to ensure that the assigned department has qualified teaching staff who are not currently overloaded with teaching duties during the corresponding time period. If, during the allocation process, there is insufficient department capacity, unavailable supervising teachers, or insufficient prior skills, the resident physician will be marked as waiting in line and given priority for re-arrangement in the next scheduling cycle.

[0010] Furthermore, the dynamic feedback sub-interface includes a workload feedback control; In the dynamic feedback sub-interface, feedback data during the execution of the rotation scheme is collected in real time, a comprehensive risk index is calculated, and the rotation scheme is adjusted, including the following steps: In response to the trigger command of the workload feedback control, a self-assessment form window is displayed to collect the workload and mental health data filled in by resident physicians. The workload includes the frequency of night shifts and the average number of surgeries participated in per day, and the mental health data includes the self-assessment results of mental health. Based on the workload and mental health data, the comprehensive risk index is calculated. When the comprehensive risk index exceeds a preset threshold, an early warning notification is automatically pushed to the management terminal, and corresponding intervention suggestions, the current skill tree, and historical rotation records are displayed in conjunction for teaching management personnel to make decisions.

[0011] Furthermore, the calculation of the comprehensive risk index based on the workload and mental health data specifically includes the following steps: Workload was quantitatively assessed by calculating the workload risk component based on the number of night shifts in the past two weeks, the frequency of departmental rotations per week, and the average number of clinical tasks per day. The mental health status was quantitatively assessed and mapped to a psychological risk component based on the standardized self-assessment scale scores submitted by resident physicians. The comprehensive risk index is obtained by summing the load risk component and the psychological risk component according to preset weights. When the comprehensive risk index exceeds the configured risk warning threshold, a tiered warning mechanism is triggered.

[0012] Furthermore, according to the intelligent management and interaction method for resident physician rotation as described in claim 1, the main interface further includes a learning task distribution control; In response to the trigger command of the control for issuing the learning task, the learning task configuration window is displayed. When a certain rotation stage is activated, the associated teaching resource package is automatically retrieved from the hospital's digital education resource library according to the skill tag corresponding to the rotation stage. This package includes the corresponding textbook chapters, operation demonstration videos, clinical guidelines, case courseware, and accompanying online test questions. The teaching resource package is pushed to the simulation training platform of the corresponding resident physician, and a learning deadline is set; The resident physicians' learning completion status, video viewing time, and test scores are recorded as evidence of their theoretical learning.

[0013] On the other hand, this application provides an intelligent management and interactive system for resident physician rotation, including the following modules: The main interface module is configured to display the main interface for intelligent management of resident physician rotations; wherein, the main interface includes a skill dependency graph management control, an intelligent rotation plan control, and a dynamic feedback control; The skill dependency graph management module is configured to: in response to a trigger command on the skill dependency graph management control, display a skill dependency graph management sub-interface, construct and maintain a skill dependency graph representing the logical dependencies between clinical operation skills, and generate and display the resident physician's skill tree based on the skill dependency graph; The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependency relationship between preceding and subsequent skills. The intelligent rotation plan module is configured to: in response to the trigger command of the intelligent rotation plan control, display the intelligent rotation plan sub-interface, collect multi-dimensional input data of resident physicians, and generate and visualize the rotation plan in combination with the skill tree, rotation cycle and teaching resource constraints, and support interactive adjustment and hard constraint verification; The dynamic feedback module is configured to: respond to the trigger command of the dynamic feedback control, display the dynamic feedback sub-interface, collect feedback data in real time during the execution of the rotation scheme, calculate the comprehensive risk index, adjust the rotation scheme, and realize closed-loop management.

[0014] The beneficial effects of this application are as follows: This application provides an intelligent interactive management method for resident physician rotations. By integrating skill dependency graph management controls, intelligent rotation plan controls, and dynamic feedback controls into the main interface, it achieves intelligent management of the entire process from skill modeling and rotation planning to execution feedback. The system first constructs a skill dependency graph in the form of a directed acyclic graph to express the pre- and post-dependent relationships between clinical operation skills, and automatically generates a visualized skill tree for each resident physician based on this graph. On this basis, combined with the resident physician's multi-dimensional input data, individual skill mastery status, preset rotation cycle, and teaching resource constraints, the system intelligently generates and visualizes rotation plans that conform to teaching logic and resource feasibility, while supporting interactive adjustments by users and real-time verification of hard constraints. During the rotation execution process, the system continuously collects actual feedback data, dynamically calculates the comprehensive risk index, and adjusts the rotation plan in a timely manner accordingly, thereby forming a closed-loop management system covering plan formulation, execution monitoring, and optimization iteration, significantly improving the scientific nature, personalization level, and operational efficiency of rotation management. This application also provides a corresponding system, the beneficial effects of which are similar to the above method, and will not be elaborated here.

[0015] Other features and advantages of this application 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 application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0016] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0017] Figure 1 This is a flowchart of the intelligent management and interaction method for resident physician rotation provided in this application; Figure 2 This is a schematic diagram of the main interface of the intelligent management interaction for resident physician rotation provided in this application; Figure 3 This is a schematic diagram of the skill dependency graph management sub-interface provided in this application; Figure 4 This is a schematic diagram of the intelligent rotation plan sub-interface provided in this application; Figure 5 This is a structural diagram of the intelligent management and interactive system for resident physician rotation provided in this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] The present application will be further described below with reference to the accompanying drawings and specific embodiments. The described embodiments should not be considered as limitations on the present application, and all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present application.

[0020] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0022] In the medical education system, standardized residency training is a crucial link in cultivating qualified clinicians, emphasizing the mastery of clinical skills through multidisciplinary rotations. Although the state has established clear rotation outlines and skill requirements, traditional management methods heavily rely on manual experience, resulting in fragmented processes, information silos, and delayed responses, making it difficult to meet the demands of modern medical education for precise, personalized, and intelligent management. While some hospitals have introduced information technology tools to assist in scheduling or recording assessments, these systems are functionally limited and lack coordination, failing to support comprehensive scientific management covering skills development, resource allocation, and individual status awareness, and particularly neglecting systematic attention to the workload and mental health of resident physicians.

[0023] Existing rotation management systems generally suffer from multiple technical deficiencies. Rotation scheduling often employs a static, uniform model, ignoring individual ability differences and leading to repetitive training or gaps in key skills. The logical dependencies between clinical skills are not explicitly modeled, making it difficult to generate learning paths that align with cognitive patterns. The planning lacks comprehensive optimization capabilities for multi-dimensional constraints such as the syllabus, departmental resources, and faculty-to-staff ratios, and does not support interactive adjustments or hard constraint verification. More significantly, the systems almost entirely fail to integrate resident physicians' workload indicators (such as the frequency of consecutive on-call shifts, night shift density, and interdepartmental handover pressure) and psychological state data (such as anxiety, burnout, and satisfaction). Mainstream tools merely mechanically assign rotation tasks, failing to perceive the physical and mental risks associated with high-intensity rotation combinations, and failing to incorporate dynamic factors such as fatigue index, sleep deprivation, or subjective stress into decision-making, resulting in plans that are detached from the trainees' actual capacity.

[0024] To address the aforementioned issues, this application proposes an intelligent management and interactive method and system for resident physician rotations. First, a skill dependency graph management module establishes clinical operational skill dependencies expressed in the form of a directed acyclic graph. Nodes represent specific skills, and directed edges represent the logical order of prerequisites and subsequent skills. Based on this, a personalized, visualized skill tree is generated for each resident physician. During the rotation plan generation phase, traditional constraints such as the teaching syllabus, rotation cycle, and departmental resources are considered to generate a rotation plan. Workload data of resident physicians (such as consecutive on-call days, night shift frequency, and intensity of interdepartmental switching) and psychological state indicators (such as anxiety level, subjective stress score, and degree of burnout) are incorporated to adjust the rotation plan, aligning it with teaching objectives and the physician's physical and mental well-being. This forms a comprehensive intelligent management system covering competency assessment, personalized arrangements, and closed-loop control, truly achieving scientific, humane, and efficient resident physician rotation management.

[0025] First, the intelligent management and interaction method for resident physician rotation provided in this application will be described in detail below with reference to the accompanying drawings.

[0026] Reference Figure 1The implementation process of the intelligent management and interaction method for resident physician rotation provided in this application embodiment includes, but is not limited to, the following steps.

[0027] Step S110: Display the main interface 100 of the intelligent management interaction for resident physician rotation.

[0028] Among them, reference Figure 2 The main interface 100 includes a skill dependency graph management control 101, an intelligent rotation plan control 102, and a dynamic feedback control 103.

[0029] In step S110, a unified user entry point and operation center are provided for the entire intelligent management and interaction method for resident physician rotations. By displaying the main interface 100, the system intuitively presents core functional modules in the form of controls, enabling teaching administrators, instructors, or resident physicians to quickly identify and access the required operation scenarios. The skill dependency graph management control 101, intelligent rotation plan control 102, and dynamic feedback control 103 integrated in the main interface 100 correspond to the three key aspects of capability modeling, planning scheduling, and execution monitoring, respectively. This structured layout not only improves human-computer interaction efficiency but also lays the foundation for the orderly invocation of subsequent sub-functions, ensuring the entire rotation management system operates collaboratively within a unified framework.

[0030] In step S120, in response to the trigger command of the skill dependency graph management control 101, the skill dependency graph management sub-interface 200 is displayed, a skill dependency graph representing the logical dependencies between clinical operation skills is constructed and maintained, and the skill tree of the resident physician is generated and displayed based on the skill dependency graph.

[0031] The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependencies between preceding and subsequent skills.

[0032] In step S120, a logical framework and individualized competency view for clinical skills training are established. When the user triggers the skills dependency graph management control 101, the system enters a sub-interface that supports the construction and maintenance of a skills dependency graph expressed in the form of a directed acyclic graph. Each node represents a specific clinical operation skill, such as endotracheal intubation or deep vein puncture, while directed edges clearly identify the prerequisite and subsequent dependencies between skills. For example, mastering aseptic technique is a prerequisite for performing puncture-like operations. Based on this graph, the system can automatically generate a personalized skills tree for each resident physician, dynamically reflecting their mastered skills and unlockable advanced paths. This not only achieves scientific control of the teaching sequence, avoiding high-risk training due to skipping basic skills, but also provides precise competency basis for personalized rotation arrangements, fundamentally ensuring the safety and effectiveness of training.

[0033] In step S130, in response to the trigger command of the intelligent rotation plan control 102, the intelligent rotation plan sub-interface 300 is displayed, multi-dimensional input data of resident physicians is collected, and the rotation plan is generated and visualized by combining the skill tree, rotation cycle and teaching resource constraints, and interactive adjustment and hard constraint verification are supported.

[0034] In step S130, the intelligent generation and flexible adjustment of the rotation plan are realized. After the user enters the intelligent rotation plan sub-interface 300, the system collects multi-dimensional input data, including basic information of resident physicians, skill mastery, theoretical learning records, teaching evaluation, rotation cycle requirements, and departmental resource capacity, and deeply integrates it with the aforementioned skill tree and teaching resource constraints. By calling the rule-based scheduling engine, the system automatically generates a rotation plan that meets all hard constraints (such as prerequisite skill attainment, departmental staffing limits, and teaching staff workload), and visualizes it in the form of Gantt charts. At the same time, this step allows users to interactively adjust the rotation period through drag-and-drop, and the system verifies the compliance of the adjustment in real time. Once a conflict is found, it immediately alerts the user, thus ensuring the scientific nature of the plan while taking into account the flexibility and operability in actual management.

[0035] In step S140, in response to the trigger command of the dynamic feedback control 103, the dynamic feedback sub-interface is displayed to collect feedback data in real time during the execution of the rotation plan, calculate the comprehensive risk index, adjust the rotation plan, and realize closed-loop management.

[0036] In step S140, a dynamic feedback and self-optimization mechanism for rotation management is constructed. When the user activates the dynamic feedback control 103, the system enters a sub-interface and continuously collects multi-source feedback data during the rotation execution process, including workload indicators reported by resident physicians (such as night shift frequency and average number of surgeries participated in per day), psychological health self-assessment results, and actual execution records from the teaching end. Based on this data, the system calculates a comprehensive risk index according to preset weights. Once the index exceeds the warning threshold, a tiered warning is automatically triggered, pushing a notification to the management end and displaying intervention suggestions, the current skill tree, and historical rotation records. More importantly, the system will dynamically adjust the original rotation plan accordingly, such as reducing the workload, changing departments, or suspending high-risk operations, thereby forming a complete closed loop from plan formulation and execution monitoring to risk intervention and plan optimization, truly achieving intelligent and humanized management centered on the growth and safety of resident physicians.

[0037] In some embodiments of this application, reference is made to Figure 3The skill dependency graph management sub-interface 200 includes a skill node addition control 201, a dependency configuration control 202, and a skill tree display area 203. In step S120, the skill dependency graph management sub-interface 200 is displayed, a skill dependency graph representing the logical dependencies between clinical operation skills is constructed and maintained, and a skill tree for resident physicians is generated and displayed based on the skill dependency graph, including the following steps.

[0038] In step S210, in response to the trigger command for adding control 201 to the skill node, the skill node editing window is displayed, and the basic information of the new skill input by the user is received. The basic information includes the skill name, the professional base, the required study hours and the qualification standards. The new skill is then added to the skill dependency graph as an independent node.

[0039] In step S210, the system provides scalable node construction capabilities for the skill dependency graph, enabling teaching administrators to dynamically introduce new clinical operation skills based on actual training needs. When a user triggers the skill node addition control 201, the system pops up a skill node editing window, guiding the user to input basic information about the new skill, including the skill name, affiliated professional base, required study hours to complete the skill, and specific achievement standards. This information not only defines the attributes of the skill itself but also provides a quantitative basis for subsequent competency assessments and rotation arrangements. After input, the system adds the skill as an independent node to the skill dependency graph. At this point, the node has not yet established dependencies with other skills and is in a pending configuration state, laying the foundation for the subsequent improvement of the logical structure.

[0040] Optionally, the achievement standards are pre-configured by the teaching management department and typically include multi-dimensional indicators such as completing a specified number of practical training sessions (e.g., at least 5 supervised central venous catheterizations), passing relevant theoretical assessments (e.g., a score of no less than 80 points), and obtaining a qualified evaluation from the supervising physician. The system automatically determines whether a skill has been completed based on these quantifiable and verifiable conditions. Simultaneously, newly added skill nodes are stored in the system as structured data, including the skill name, unique identifier, affiliated specialty base (using the national standardized residency training specialty directory code), required training hours, and achievement standard fields, ensuring that subsequent dependency configuration and competency assessment are based on verifiable data.

[0041] In step S220, in response to the trigger command of the dependency configuration control 202, the dependency configuration window is displayed. A directed dependency edge is established between two skill nodes by dragging or selecting. During the establishment process, it is checked whether a loop is formed. If a loop is detected, saving is prohibited and the user is prompted to correct it, so as to ensure that the skill dependency graph always maintains a directed acyclic structure.

[0042] In step S220, the structural rigor and logical rationality of the skill dependency graph are ensured. After the user enters the configuration window through the dependency configuration control 202, they can establish directed dependency edges between two existing skill nodes by dragging or selecting from a dropdown menu. This clearly expresses the logical relationship that a skill can only be learned after its prerequisite skills have been mastered. During this process, the system executes a loop detection algorithm in real time. If it detects that the established dependency relationship will lead to a circular dependency in the graph (e.g., skill A depends on skill B, and skill B indirectly depends on skill A), the system will immediately prevent the save operation and issue a correction prompt to the user. This mechanism effectively ensures that the skill dependency graph always maintains the mathematical characteristics of a directed acyclic graph, thereby preventing the teaching path from falling into a logical dead loop and ensuring that the learning path for resident physicians has a clear sequence and is executable.

[0043] Optionally, when a user attempts to establish a directed edge from skill A to skill B, the system will perform a depth-first search (DFS) starting from skill B. If skill A is revisited during the traversal, a cycle is identified. Alternatively, an incremental topological sorting method can be used, attempting to sort the local subgraph after each new edge is added. If the sorting fails, it indicates the introduction of a circular dependency. Once a cycle is detected, the system immediately blocks the save operation and displays a message on the interface stating, "The selected dependency will cause a skill logic loop; please adjust the prerequisite relationships." This ensures that the skill dependency graph always maintains the mathematical properties of a directed acyclic graph, guaranteeing the logical rationality and executability of the teaching path.

[0044] Step S230: Based on the skill dependency graph, generate a skill tree for each resident physician, which includes automatically matching the corresponding prerequisite skills and subsequent skill paths according to the skill items that the resident physician has completed.

[0045] In step S230, the general skill dependency graph is individually mapped onto each resident physician, forming a personalized skill development path that matches their current competency level. Based on the skill items that the resident physician has completed and passed assessments, the system automatically traces upwards through the skill dependency graph to verify its completeness and downwards to deduce all unlockable subsequent skills, thereby generating a skill tree reflecting their current competency boundaries and future growth direction. This skill tree not only identifies the skill statuses of those already mastered, those needing consolidation, and those that can be advanced, but also implicitly contains the optimal learning sequence, providing a precise competency profile for the formulation of subsequent intelligent rotation plans, enabling training arrangements to truly achieve individualized instruction and gradual progression.

[0046] Step S240: Display the skill tree of the selected resident physician in the skill tree display area 203.

[0047] In step S240, the system uses visualization to visually present the individual resident physician's competency structure and development potential, improving the transparency and decision-making efficiency of teaching management. In the skill tree display area 203, the system clearly displays the selected resident physician's skill tree in the form of a tree diagram or hierarchical diagram. Different colors or icons distinguish the mastery status of skills, and the branching structure reflects the dependency logic between skills. Users can expand or collapse subtrees to focus on specific professional areas. This visualization not only facilitates resident physicians' self-awareness of their current learning progress but also helps instructors quickly identify trainees' weaknesses. Furthermore, it provides a direct basis for teaching management departments to evaluate overall training effectiveness and adjust resource allocation, thereby strengthening the observability, traceability, and interventionability of the skills development process.

[0048] In some embodiments of this application, step S240 involves generating a skill tree for each resident physician based on a skill dependency graph, including the following steps.

[0049] Step S310: Obtain the set of skills that the current resident physician has completed.

[0050] Optionally, the set of completed skill items comes from the system's internal training record database. Each record contains a unique identifier for the skill, completion time, certification status (such as "passed the assessment"), assessment method, and scoring result. The skill is considered completed and included in the set only when the certification status is valid.

[0051] In step S310, authentic and accurate starting data is provided for constructing personalized skill trees for resident physicians. The system first retrieves a set of skill items that the current resident physician has completed and passed assessments from the training record database or assessment system. These items represent the competency units that the resident physician has mastered and certified in clinical practice. This set forms the basis for all subsequent reasoning and path deduction, ensuring that the generated skill tree closely matches the resident physician's actual competency level, rather than being based on assumptions or generic templates. This guarantees the practical feasibility of the skill development path and the targeted nature of the teaching guidance.

[0052] Step S320: In the skill dependency graph, mark the nodes corresponding to the completed skill items as mastered.

[0053] In step S320, the resident physician's actual competency status is mapped onto the structure of a general skill dependency graph. The system locates nodes corresponding to completed skill items in the skill dependency graph and explicitly marks their status as "mastered." This marking is not only a visual or logical identifier but also a triggering condition for subsequent dependency reasoning. Through this status labeling, the system can clearly distinguish which skills are the foundation of current competency and which are still inactive, thus providing a reliable basis for automatically identifying the next learnable skill. This allows the entire skill tree generation process to be built upon dynamic, real-time competency cognition.

[0054] Step S330: Starting from all mastered nodes, traverse the skill dependency graph, identify all unlocked skill nodes whose prerequisite skills have been marked as mastered, and add the unlocked skill nodes to the unlockable skill set.

[0055] In step S330, a dependency-based intelligent skill unlocking mechanism is implemented. Starting with all mastered nodes, the system performs a depth-first traversal of the skill dependency graph, checking each ununlocked skill node to see if it meets the condition that all its prerequisite skills have been mastered. For skill nodes where all dependencies have been marked as mastered, the system identifies them as currently learnable advanced skills and includes them in the unlockable skill set. This process simulates the natural learning pattern of "basic before advanced" in clinical teaching, ensuring that resident physicians are not recommended advanced procedures before they possess the necessary prerequisites, thereby improving training efficiency while ensuring patient safety.

[0056] Optionally, the system employs a breadth-first search (BFS) strategy, starting from all mastered nodes and examining their direct successor nodes layer by layer. For each unlocked node, the system quickly retrieves all its direct predecessor skills using pre-stored incoming edge indices and verifies whether all these predecessor nodes have been marked as "mastered." Only when all conditions are met is the node added to the unlockable skill set. Simultaneously, unlockable skills are marked as "learnable" or "awaiting mastery" in the skill tree, rather than "mastered," to avoid logical confusion.

[0057] Step S340: Add the skill nodes from the unlockable skill set to the resident physician's skill tree and update the mastered status.

[0058] In step S340, the reasoning results are transformed into actual content in the skill tree and their status is updated synchronously. The system formally adds all skill nodes from the unlockable skill set identified in the previous step to the resident physician's skill tree and updates their status to learnable or pending mastery. In some embodiments, these nodes may also be pre-displayed as potentially mastered skills. This operation not only enriches the structure of the skill tree but also dynamically reflects the resident physician's current learning stage and future development direction. By continuously integrating newly unlocked skills, the skill tree gradually evolves from its initial state to a complete competency map, forming a personalized competency model that grows continuously with the training process.

[0059] Step S350: Repeat the traversal and recognition steps until the set of unlockable skills is empty, thereby generating a complete skill tree.

[0060] In step S350, the completeness and convergence of the skill tree generation process are ensured. Since skill dependencies may involve multiple layers of nesting, unlocking certain higher-order skills depends on the prior unlocking of intermediate skills. Therefore, a single traversal may not discover all learnable nodes. The system progressively unlocks skills by repeatedly executing the traversal and identification process starting from already mastered nodes, until no new unlockable skills are found in a certain round of traversal, at which point the set of unlockable skills is empty. This iterative mechanism ensures that the skill tree ultimately covers all skill nodes achievable based on current capabilities, forming a logically complete, structurally closed, and fully adaptable skill development panorama tailored to the current level of resident physicians, providing a solid capability basis for subsequent rotation planning.

[0061] Furthermore, since the skill dependency graph is a directed acyclic graph (DAG), if no new unlockable skills are found after each round of traversal, the system determines that the unlockable set is empty and the iteration terminates. This mechanism essentially performs reachability closure deduction on the graph under the current ability state, ensuring that the final generated skill tree covers all logically upgradable skill nodes. To support efficient traversal, the skill dependency graph is stored in the system as an adjacency list, with each node maintaining its outgoing (successor) and incoming (preceding) edge lists, thus supporting fast state queries and dependency verification.

[0062] In some embodiments of this application, the system retrieves the set of completed skill items for the current resident physician from an internal training record database. Each record includes a skill ID, completion time, certification status, and evaluation result. Based on this, the skill tree is not statically pre-stored but rather a logical view generated in real-time as needed. The generation process starts with all mastered nodes and uses a breadth-first search (BFS) strategy to traverse the skill dependency graph layer by layer, identifying unlocked nodes where all prerequisite skills have been satisfied and including them in the learnable scope. This process supports dynamic updates—when a resident physician completes a new skill and it is confirmed by the system, their skill tree will be automatically recalculated on the next visit, ensuring that it always reflects the latest capability status.

[0063] In some embodiments of this application, the skill dependency graph management sub-interface 200 further includes a version management control 204. In response to a trigger command on the version management control 204, a version management window is displayed, supporting saving the current skill dependency graph version, publishing to the production environment, rolling back to the previous version, and branch management and independent maintenance of the skill dependency graph according to different professional bases.

[0064] Specifically, the version management function provides full lifecycle version control and multi-professional collaborative management capabilities for the skill dependency graph, ensuring the stability, traceability, and flexibility of the teaching system during its continuous evolution. When a user triggers the version management control 204, the system pops up a version management window, allowing teaching administrators to completely save the currently edited skill dependency graph in a timestamp and version number format, forming an immutable historical snapshot. It also supports officially releasing the approved graph version to the production environment, making it effective and usable for subsequent skill tree generation and rotation planning for resident physicians. If problems are found in the new version during actual application, the system also provides a one-click rollback function to quickly restore to the previous stable version, minimizing the risk of teaching arrangements being interrupted due to configuration errors.

[0065] More importantly, considering the significant differences in clinical skill requirements among different professional bases (such as internal medicine, surgery, anesthesiology, etc.), the system supports branch management of skill dependency maps by professional base. Each specialty can independently create, edit, and maintain its own exclusive skill dependency structure without interfering with each other. This enables highly customized teaching standard management under a unified platform architecture, taking into account both standardization and professional specificity.

[0066] In some embodiments of this application, reference is made to Figure 4The intelligent rotation plan sub-interface 300 includes a data import control 301, an operation video upload control 302, a rotation plan generation control 303, and a plan adjustment control 304. In step S130, the intelligent rotation plan sub-interface 300 collects multi-dimensional input data from resident physicians, combines the skill tree, rotation cycle, and teaching resource constraints, generates and visualizes the rotation plan, and supports interactive adjustment and hard constraint verification, including the following steps.

[0067] In step S410, in response to the trigger command of the data import control 301, the data import window is displayed, and the electronic medical record system, simulation training platform or mobile terminal scoring interface is configured and connected to obtain the resident physician's theoretical learning status, teaching score and personal basic information.

[0068] In step S410, multi-source heterogeneous data channels are established to provide a comprehensive, real-time, and structured profile of resident physician competence for the generation of the intelligent rotation plan. When the user triggers the data import control 301, the system pops up a data import window, supporting configuration and access to external data sources such as the hospital's internal electronic medical record system, simulation training platform, or mobile scoring interface. Through these interfaces, the system automatically obtains the resident physician's theoretical learning completion status (such as course hours and assessment scores), the instructor's scoring records of their clinical performance, and personal basic information (such as name, specialization, and current rotation stage). This multi-dimensional data not only enriches the dimensions of resident physician competence assessment but also lays the data foundation for the subsequent dynamic updating of skill trees and the accurate matching of rotation plans, avoiding the lag and subjective bias caused by relying on manual data entry, thereby improving the objectivity and scientific nature of the entire rotation arrangement.

[0069] Furthermore, the system interfaces with existing hospital electronic medical record systems, simulation training platforms, or mobile teaching applications via standard interface protocols (such as HTTP RESTful API). If the hospital lacks a structured scoring system, it also supports manual batch import of teaching evaluation sheets in Excel or CSV format. The system provides a field mapping configuration interface, allowing administrators to map columns such as "Student ID," "Course Name," and "Assessment Score" from external systems to the internal data model, completing data normalization without requiring custom middleware development; it only relies on the hospital's standard data export and interface access capabilities.

[0070] In step S420, in response to the trigger command of the operation video upload control 302, a video acquisition window is displayed to receive the clinical operation video uploaded by the resident physician or teaching staff, perform posture estimation and instrument trajectory recognition, and generate a structured skill score.

[0071] In step S420, an AI-based objective skills assessment method is introduced, transforming traditional clinical operation evaluations that rely on subjective scoring into quantifiable and traceable structured data. After the user triggers the operation video upload control 302, the system displays a video capture window, allowing resident physicians or instructors to upload recordings of specific clinical operation procedures, such as cardiopulmonary resuscitation or lumbar puncture. The system then calls upon a built-in computer vision model to perform posture estimation on the video to analyze the standardization of the operator's limb movements, and combines instrument trajectory recognition technology to track the spatial motion path of key tools. Based on a preset operation standard template, the system automatically generates a structured skills score that includes dimensions such as action completeness, timing compliance, and instrument usage accuracy. This score not only serves as important evidence of skill mastery but can also be used to correct or supplement ability assessments in the skills tree, making rotation arrangements closer to actual operational levels and significantly improving training quality and patient safety.

[0072] Optionally, after receiving clinical operation videos uploaded by users, the system forwards them to the open API of a third-party general multimodal big data model. The call includes prompts such as, "Please analyze this endotracheal intubation operation video, assess the operator's hand movements and whether the sequence of instrument use conforms to guidelines, and provide a comprehensive skill score from 0 to 100, listing the main deduction points." The big data model returns structured JSON results (e.g., {"score": 82, "issues": ["Patient position not confirmed", "Laryngoscope holding angle deviation"]}), which the system parses and stores in the skill score database. The entire process does not deploy a local AI model; it only acts as a client for calling the existing big data model's intelligent agent, significantly reducing the hospital's computing power and operational barriers.

[0073] In step S430, in response to the trigger command of the rotation scheme generation control 303, based on the skill tree, skill score, rotation cycle limit, department capacity limit and teaching resource status, the rule-based scheduling engine is invoked to generate a rotation scheme that meets the hard constraints of the prerequisite skill achievement, and the scheme is visualized in the form of a Gantt chart.

[0074] In step S430, intelligent collaborative scheduling among teaching resources, individual capabilities, and institutional constraints is achieved, automatically generating a scientifically sound and feasible personalized rotation plan. When the user clicks the rotation plan generation control 303, the system comprehensively considers multiple factors, including the resident physician's current skill tree status, skill scores generated from video analysis or teaching evaluation, the prescribed rotation cycle limit, the receiving capacity limits of each department, and the available resource status of the teaching staff, and invokes a rule-based scheduling engine. This engine has built-in hard constraint rules, such as requiring prior mastery of pre-requisite skills like aseptic technique and intravenous puncture in a certain rotation department; if these requirements are not met, the rotation is prohibited. Under the premise of satisfying all hard constraints, the engine optimizes the allocation of rotation time slots, ultimately outputting a complete rotation plan, which is visualized on the interface in the form of a Gantt chart, clearly showing the department to which each resident physician belongs at different time slots and their skill advancement path, greatly improving the efficiency and transparency of plan formulation.

[0075] Optionally, the scheduling engine's rule logic is configured by teaching administrators via natural language or dropdown options on the interface (e.g., "CPR and intubation must be completed before entering the ICU, and the score must be ≥80"). The system automatically converts this into internal conditional statements. The scheduling process employs deterministic algorithms (e.g., greedy matching + priority queue) rather than complex optimization solutions. The Gantt chart is rendered using an open-source front-end library, and the data comes from the scheduling result table, requiring no special visualization engine.

[0076] In step S440, in response to the trigger command of the scheme adjustment control 304, the user is allowed to drag and drop to adjust the rotation period, and the scheme after adjustment is checked in real time to see if it violates the hard constraints. When there is a conflict, a prompt message will pop up to ensure the compliance and executability of the rotation plan.

[0077] In step S440, teaching administrators are given the ability to manually intervene based on automated generation, while ensuring that any adjustments do not violate core teaching rules and resource constraints. When a user uses the scheme adjustment control 304 to drag and modify the rotation period in the Gantt chart, the system immediately activates a real-time verification mechanism to dynamically check whether the adjusted arrangement violates hard constraints, such as whether it assigns resident physicians to departments where they have not yet mastered the prerequisite skills, whether it exceeds the maximum number of patients a department can admit, or whether it conflicts with the scheduling of supervising teachers. Once a violation is detected, the system immediately displays a prompt message, clearly indicating the type of conflict and the relevant constraint rules, and prevents unauthorized saving.

[0078] Optionally, when a user drags the carousel, the front-end immediately calls the back-end lightweight validation interface. This interface only checks three hard constraints: whether the target department is overcrowded during the current time period, whether the physician has met all prerequisite skills, and whether there are available teaching staff during the corresponding time period (teacher workload data comes from the scheduling table). The validation logic is a simple conditional judgment and does not involve complex reasoning. Conflict warning messages directly quote preset text (such as "Department capacity is full").

[0079] In some embodiments of this application, step S430 involves invoking a rule-based scheduling engine to generate a round-robin scheme, including the following steps.

[0080] Step S510: Based on the skill tree and skill scores, determine the set of rotation departments that each resident physician can currently enter. A department is only included in the set of rotation departments if the skill scores corresponding to all the prerequisite skills required by a department reach the preset threshold.

[0081] In step S510, a departmental selection logic based on competency achievement as the core admission mechanism is constructed to ensure that resident physicians are only assigned to rotation environments where their current skill level is sufficient for safe performance. The system comprehensively analyzes each resident physician's skill tree structure and corresponding skill scores, comparing each item against the prerequisite skill requirements set by the target department. Only when the physician's scores on all required skills reach or exceed the preset thresholds (e.g., aseptic technique score not lower than 85 points, passing basic life support assessment, etc.) will that department be included in the set of rotation departments to which the physician can enter. This mechanism fundamentally eliminates the possibility of being assigned to high-risk operations due to insufficient skill preparation, embedding patient safety and training quality into the rotation admission rules, and realizing a paradigm shift from "rotation by time" to "rotation by competency advancement."

[0082] Step S520: Based on the preset rotation cycle and the maximum number of people each department can receive, allocate department rotation time slots to resident physicians according to the first-come, first-served or preset rotation priority order.

[0083] In step S520, under the premise of meeting competency access requirements, the system coordinates the time and capacity constraints of teaching resources to achieve a fair and orderly allocation of rotation slots. Based on the total rotation cycle stipulated by the state or hospital, and combined with the maximum number of patients each department can accept within a specific time period (e.g., the cardiology department can accept a maximum of 6 resident physicians per month), the system schedules rotations according to a predetermined strategy. This strategy can be configured as a first-come, first-served basis (based on the order of registration or completion of pre-training) or a preset rotation priority (e.g., priority for core departments, preferential treatment for shortage specialties, etc.). Through this rule-driven allocation method, the system ensures teaching order while taking into account the structural requirements of different professional training programs, avoiding resource congestion or emptying of key departments, and improving the overall operational efficiency and planning of the training system.

[0084] Step S530: Synchronously match teaching resources to ensure that the assigned department has qualified teaching staff within the corresponding time period and that the current teaching workload is not exceeded.

[0085] In step S530, teaching human resources are incorporated into the rotation scheduling closed loop to ensure that each rotation arrangement not only has departmental capacity but also qualified instructors. After assigning a specific department and time slot to the resident physician, the system simultaneously queries the scheduling of instructors for that department during the corresponding time slot to verify whether there are available instructors with the corresponding professional qualifications (such as senior professional titles or specialized certifications) and whose current teaching load does not exceed the preset limit (e.g., each instructor can only supervise a maximum of 3 trainees at a time). If a match is found, the teaching relationship is locked; if no suitable instructor is available, it is considered a resource conflict. This step effectively prevents the phenomenon of "no one to teach a position" or "overloaded teaching staff with multiple trainees," ensuring the quality of clinical teaching and the effectiveness of teacher-student interaction, making rotation not only a job experience but also a process of skill development with guidance and feedback.

[0086] In step S540, if there is insufficient department capacity, unavailable teaching staff, or unmet prerequisite skills during the allocation process, the resident physician is marked as waiting in line and given priority for re-arrangement in the next scheduling cycle.

[0087] In step S540, a flexible scheduling and dynamic retry mechanism is established to improve the system's robustness and service continuity under complex real-world constraints. When an immediate arrangement cannot be completed during the allocation process due to reasons such as department capacity being full, teaching staff being unavailable, or resident physicians not fully meeting the prerequisite skill requirements, the system will not directly abandon or report an error. Instead, it will automatically mark the resident physician as waiting in the queue and record the specific conditions that were not met.

[0088] In the next scheduling cycle (when the next monthly or next scheduling window opens), the system will prioritize reassigning these personnel, and will immediately fill the vacancies once relevant resources become available or their skills meet the requirements. This queuing and retry mechanism avoids the tediousness of repeated manual intervention while ensuring that the training process of each resident physician is not delayed indefinitely. It reflects the system's balance between rigid rules and flexible management, enhancing the adaptability and humanization of the overall rotation management system.

[0089] In some embodiments of this application, taking the scenario of the scheduling engine generating a rotation plan as an example, the system first configures a structured admission rule table for each rotation department. For example, the cardiology department requires "ECG interpretation skill score ≥ 80 points" and "basic life support score ≥ 85 points". Each rule is stored in the database in the form of (department ID, skill ID, and threshold). When step S510 is executed, the system reads the skill score data of a resident physician and compares it item by item with the admission list of the cardiology department. If both scores meet the requirements, the cardiology department is added to the set of rotation departments that the physician can currently enter; otherwise, it is excluded.

[0090] After entering step S520, the system sorts the physicians to be assigned rotation slots according to the rotation priority preset by the teaching management department (e.g., "Emergency Department > ICU > Cardiology"), and assigns them rotation slots in sequence. Taking Cardiology as an example, the system queries the number of physicians already assigned to the department this month (initially 0, with a maximum of 6). If there are currently 3 physicians and available slots, the assignment is allowed, and the number is updated to 4. This process is completed through simple SQL queries and counter updates to ensure that the capacity is not exceeded.

[0091] In step S530, the system synchronously matches teaching resources. Each teacher registers their qualifications (e.g., "Dr. Li: Can teach cardiology and respiratory medicine") and maximum teaching load (e.g., "Maximum 3 students at a time") in the system. The system queries all qualified cardiology teachers within the target time period and counts their current number of students. If Dr. Zhang has already taught 2 students, he can accept 1 more, and the system binds him as Dr. Zhang's teaching teacher. If all qualified teachers are at full capacity, the system determines that the teaching resources are unavailable.

[0092] If any of the above steps fail (e.g., skills not met, department full, or no available teachers), the system inserts the physician's record into the "Rotation Allocation Queue" table in step S540. Fields include physician ID, reason for failure, first attempt time, and number of retries. At midnight on the 1st of each month, the system automatically starts a new round of scheduling via a scheduled task (such as Linux cron or a database job), prioritizing records in this queue. If the conditions are met (e.g., newly completed skills training or a teacher releasing a spot), the physician is assigned and removed from the queue; otherwise, the record is retained and the number of retries is accumulated.

[0093] In some embodiments of this application, the dynamic feedback sub-interface includes a workload feedback control. In step S140, the dynamic feedback sub-interface collects feedback data in real time during the execution of the rotation plan, calculates a comprehensive risk index, and adjusts the rotation plan, including the following steps.

[0094] Step S610: In response to the trigger command of the workload feedback control, the self-assessment form window is displayed, and the workload and mental health data filled in by the resident physicians are collected. The workload includes the frequency of night shifts and the average number of surgeries participated in per day, and the mental health data includes the self-assessment results of mental health.

[0095] In step S610, a direct feedback channel is established between the resident physician's subjective experience and the system's objective management, enabling the quantification of their physical and mental workload during rotations and incorporating it into the intelligent control system. When teaching administrators or the resident physician triggers the workload feedback control, the system pops up a self-assessment form window, guiding the resident physician to regularly report their actual workload and psychological feelings during their current rotation phase. Workload indicators specifically include statistically available objective data such as weekly night shift frequency and average number of surgeries performed per day. Psychological health data is collected through standardized self-assessment scales (such as a simplified version of PHQ-4 or GAD-2) to collect self-assessment results regarding anxiety, depression, or burnout tendencies. This data is submitted and stored in a structured format, not only reflecting the individual's true stress level but also providing crucial input for subsequent risk identification, shifting rotation management from simple task assignment to a comprehensive training model that considers both occupational health and humanistic care.

[0096] Step S620: Based on workload and mental health data, calculate the comprehensive risk index. When the comprehensive risk index exceeds the preset threshold, automatically push an early warning notification to the management terminal and display the corresponding intervention suggestions, current skill tree and historical rotation records for teaching management personnel to make decisions.

[0097] In step S620, subjective feedback is transformed into actionable management intervention signals, realizing a closed-loop control mechanism from passive response to proactive early warning. The system calculates a comprehensive risk index between 0 and 100 based on a pre-set weighted model, weighting the workload and mental health data submitted by resident physicians. For example, a combination of excessively high night shift frequency, intensive surgical involvement, and low mental health scores will significantly increase this index. Once the index exceeds the safety threshold set by the teaching management department (e.g., 75 points), the system immediately pushes a prominent early warning notification to the management end, simultaneously displaying three key pieces of information: standardized intervention recommendations for the current risk type (e.g., "suggest suspending night shifts for two weeks" and "arrange psychological support talks"), the physician's latest skill tree status (reflecting whether their skill development is hindered), and complete historical rotation records (facilitating the tracking of workload accumulation). This mechanism enables teaching management personnel to intervene promptly before problems worsen, dynamically adjusting subsequent rotation arrangements, thereby protecting the physical and mental health of resident physicians while ensuring training quality, truly achieving an intelligent, humane, and sustainable closed-loop medical talent training system.

[0098] In some embodiments of this application, step S620 involves calculating a comprehensive risk index based on workload and mental health data, specifically including the following steps.

[0099] Step S710: Quantitatively assess the workload by calculating the workload risk component based on the number of night shifts in the past two weeks, the frequency of departmental rotations per week, and the average number of clinical tasks per day.

[0100] In step S710, the objective workload experienced by resident physicians during rotations is transformed into calculable and comparable numerical indicators, providing a solid factual basis for risk assessment. The system focuses on key workload dimensions over the past two weeks, including the number of night shifts (reflecting the degree of disruption to physiological rhythms), the frequency of rotational department changes within a single week (reflecting adaptation stress and the risk of interruption of learning continuity), and the average daily clinical tasks (such as the number of patients seen, the amount of procedures performed, etc., characterizing the level of daily cognitive and physical exertion). A scoring coefficient is assigned to each item according to teaching management rules; for example, 5 points are added for each additional night shift, 8 points for more than 2 departmental changes, and 3 points for exceeding the baseline value for the average daily tasks. These scores are accumulated to form a workload risk component, realistically depicting the actual stress state of an individual in a high-intensity clinical environment, avoiding judgments of fatigue levels based solely on subjective impressions.

[0101] Step S720: Quantitatively assess mental health status by mapping the scores of standardized self-assessment scales submitted by resident physicians to a psychological risk component.

[0102] In step S720, the mental health status of resident physicians is transformed from vague perceptions into structured, traceable quantitative parameters. The system uses standardized self-assessment scales (such as the Patient Health Questionnaire-4 or the Generalized Anxiety Disorder Scale-2) that have been widely validated in medical education. After resident physicians complete the scales online regularly, the system converts their total scores into a psychological risk component according to a preset mapping rule; for example, a total score of 0 to 2 indicates low risk (corresponding to a score of 10), 3 to 5 indicates medium risk (corresponding to a score of 40), and 6 or above indicates high risk (corresponding to a score of 80). This component not only reflects the current level of emotional distress but also implies early signs of burnout or adjustment disorders, making mental health no longer a blind spot in rotation management but an important dimension that can be monitored and intervened.

[0103] Step S730: The load risk component and the psychological risk component are weighted and summed according to preset weights to obtain the comprehensive risk index.

[0104] In step S730, objective workload and subjective psychological state are integrated to construct a unified and balanced comprehensive risk measurement system. The system weights and sums the workload risk component and psychological risk component obtained in the first two steps according to the weights pre-configured by the teaching management department, for example, workload accounts for 60% and psychological risk accounts for 40%, to reflect the medical talent training concept of "interactive influence of mind and body." If a physician's workload component is 70 and psychological risk component is 80, then the comprehensive risk index is 70 multiplied by 0.6 plus 80 multiplied by 0.4, which equals 74. This index does not overly rely on a single indicator, but can sensitively reflect the overall risk level under multiple pressures, providing a scientific and transparent numerical basis for subsequent early warning decisions.

[0105] Step S740: When the comprehensive risk index exceeds the configured risk warning threshold, the tiered warning mechanism is triggered.

[0106] In step S740, the risk response is made more precise and tiered to ensure that intervention measures match the severity of the risk. When the comprehensive risk index exceeds the warning thresholds configured in the system (e.g., Level 1 warning threshold 60, Level 2 75, Level 3 90), the system automatically triggers the corresponding tiered warning mechanism; for example, at Level 1, only a reminder is recorded; at Level 2, a notification is sent to the teaching secretary and a suggestion to adjust the schedule is made; at Level 3, a high-priority alarm is immediately sent to the teaching director, and high-risk rotation arrangements are forcibly suspended. Each level of warning is associated with a preset intervention process and responsible person, making the management response both timely and orderly, effectively preventing medical errors, training interruptions, or physical and mental health crises caused by accumulated stress, and truly realizing an intelligent closed-loop protection system centered on resident physicians.

[0107] In some embodiments of this application, the main interface 100 also includes a learning task distribution control 104 to implement the following steps.

[0108] In step S810, in response to the trigger command of the learning task distribution control 104, the learning task configuration window is displayed. When a certain rotation stage is activated, the associated teaching resource package is automatically retrieved from the hospital's digital education resource library according to the skill tag corresponding to the rotation stage. This package includes the corresponding textbook chapters, operation demonstration videos, clinical guidelines, case courseware, and accompanying online test questions.

[0109] In step S810, intelligent association between rotation stages and teaching resources is implemented to ensure that resident physicians receive accurate and timely theoretical support before entering specific clinical positions. When teaching administrators or the system automatically triggers the learning task distribution control 104, the system pops up a learning task configuration window. When a certain rotation stage is activated (e.g., a resident physician is about to enter a cardiology rotation), the system automatically extracts the skill tags bound to that stage (such as "ECG interpretation" and "acute heart failure management"). Subsequently, based on these tags, the system intelligently retrieves and aggregates relevant teaching resource packages from the hospital's existing digital education resource library. The content covers designated chapters of authoritative textbooks, standardized operation demonstration videos, the latest version of clinical diagnosis and treatment guidelines, typical case analysis courseware, and accompanying online test questions. This mechanism dynamically integrates scattered teaching materials according to ability requirements, avoiding physicians blindly searching and significantly improving the relevance and efficiency of learning.

[0110] Step S820: Push the teaching resource package to the simulation training platform of the corresponding resident physician and set the learning deadline.

[0111] In step S820, a closed-loop channel is established between teaching resources and training execution, transforming passive learning into active, task-driven learning. After matching resource packages, the system automatically pushes them to the simulation training platform (such as a hospital-built online learning system or a partnered medical education platform) under the corresponding resident physician's personal account, and clearly sets a learning deadline in the task, which is usually linked to the rotation start time (e.g., requiring completion 3 days before rotation entry). Through mandatory task assignment and time constraints, the system ensures that theoretical preparation precedes clinical practice, reinforcing the safety standard of "learn before doing," laying a cognitive foundation for subsequent skills operations, and reducing the burden on instructors to repeatedly explain basic content, allowing clinical teaching to focus on high-level guidance and feedback.

[0112] Step S830: Record the resident physician's learning completion status, video viewing time, and test score as the resident physician's theoretical learning situation.

[0113] In step S830, a quantifiable and traceable theoretical learning assessment system is constructed to provide objective data support for competency profiling. The system records the entire process of resident physicians' use of teaching resources, including whether all learning content has been completed, the actual viewing time of demonstration videos (to prevent fast-forwarding or idling), and the final score and answer details of online tests. This data is stored in a structured format in the training file, serving as a core component of their theoretical learning. It is used not only to determine whether the prerequisites for entering the next skills training are met (e.g., a test score ≥80 is required to unlock practical operation access), but also for instructors to reference during interviews or feedback sessions to identify knowledge gaps. By digitizing the learning process and quantifying the results, the system achieves refined management from "whether the learning was done" to "how well it was done," effectively supporting a competency-oriented resident physician training model.

[0114] Secondly, refer to Figure 5 This application provides an intelligent management and interactive system for resident physician rotation, comprising the following modules: The main interface module is configured to display the main interface for the intelligent management of resident physician rotations. This main interface includes controls for skill dependency graph management, intelligent rotation planning, and dynamic feedback.

[0115] The skill dependency graph management module is configured to: respond to the trigger command of the skill dependency graph management control, display the skill dependency graph management sub-interface, construct and maintain a skill dependency graph representing the logical dependencies between clinical operation skills, and generate and display the resident physician's skill tree based on the skill dependency graph.

[0116] The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependencies between preceding and subsequent skills.

[0117] The intelligent rotation plan module is configured to: respond to the trigger command of the intelligent rotation plan control, display the intelligent rotation plan sub-interface, collect multi-dimensional input data of resident physicians, combine the skill tree, rotation cycle and teaching resource constraints, generate and visualize the rotation plan, and support interactive adjustment and hard constraint verification.

[0118] The dynamic feedback module is configured to: respond to the trigger command of the dynamic feedback control, display the dynamic feedback sub-interface, collect feedback data in real time during the execution of the rotation plan, calculate the comprehensive risk index, adjust the rotation plan, and realize closed-loop management.

[0119] In summary, the intelligent management and interaction method and system for resident physician rotation provided in this application have the following technical effects.

[0120] This application constructs a directed acyclic graph model centered on a skill dependency graph to structure the logical dependencies between clinical operational skills and dynamically generates personalized skill trees for each resident physician. This achieves a training model based on "ability advancement" rather than "time rotation," effectively avoiding clinical risks caused by a lack of prerequisite skills. The system integrates multi-dimensional data such as theoretical learning records, teaching evaluations, operational video analysis, and personal basic information. Combined with hard constraints such as departmental capacity, rotation cycle, and teaching resources, it uses a rule-based scheduling engine to automatically generate compliant rotation plans and supports interactive adjustments and real-time conflict verification, significantly improving scheduling efficiency, resource utilization, and plan executability.

[0121] Meanwhile, the system introduces a dynamic feedback and early warning mechanism. By collecting workload and mental health data through resident physician self-assessment, a comprehensive risk index is quantified and calculated. When the index exceeds the limit, a tiered early warning is automatically triggered, intervention suggestions are pushed, and the system displays the resident physician's skill tree and historical rotation records, enabling proactive monitoring and closed-loop intervention of their physical and mental state. Furthermore, when a rotation phase is activated, the system can automatically match and distribute a teaching resource package containing textbooks, videos, guidelines, case studies, and test questions from the hospital's digital education resource library based on skill tags. Deadlines are set, and learning behavior is recorded throughout the process to ensure a close connection between theory and practice. In summary, this application constructs a data-driven, competency-oriented, safe, efficient, and humanistic intelligent training system for resident physicians.

[0122] It should be noted that in all specific embodiments of this application, all data processing activities related to user identity or personal characteristics, such as user information, user behavior data, historical data, and location information, will be conducted in accordance with the principles of legality, legitimacy, and necessity. All data collection, use, storage, and processing will be subject to compliance with applicable national and regional laws, regulations, and industry standards, and informed consent from users will be obtained in a clear and explicit manner before processing. For the processing of sensitive personal information, separate consent from users will be obtained through prominent means such as pop-up prompts and independent confirmation pages. If any processing conflicts with laws and regulations, the laws and regulations will prevail, and necessary data processing will only be carried out within the scope permitted by laws and regulations, ensuring that all data-based applications, analyses, and technical implementations are conducted within the scope permitted by laws and regulations.

[0123] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this application are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0124] Furthermore, although this application is described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding this application. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of ordinary skill of an engineer. Therefore, those skilled in the art can implement the application set forth in the claims using ordinary skill. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of this application, which is determined by the full scope of the appended claims and their equivalents.

[0125] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several programs to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0126] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable programs for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, a program execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can retrieve and execute a program from or in conjunction with such a program execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit a program for use by or in conjunction with a program execution system, apparatus, or device.

[0127] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Additionally, computer-readable media can even be paper or other suitable media on which programs can be printed, for example, by optically scanning the paper or other media, then editing, interpreting, or, if necessary, processing it in a suitable manner to obtain the program electronically, and then storing it in computer memory.

[0128] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable program execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0129] In the foregoing description of this specification, the reference to terms such as "one embodiment / implementation," "another embodiment / implementation," or "certain embodiments / implementations," etc., indicates that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in an embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0130] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0131] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A smart management and interactive method for resident physician rotation, characterized in that, Includes the following steps: The main interface displays the intelligent management interaction interface for resident physician rotations; wherein, the main interface includes a skill dependency graph management control, an intelligent rotation plan control, and a dynamic feedback control; In response to the trigger command of the skill dependency graph management control, the skill dependency graph management sub-interface is displayed, a skill dependency graph representing the logical dependencies between clinical operation skills is constructed and maintained, and a skill tree of the resident physician is generated and displayed based on the skill dependency graph. The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependency relationship between preceding and subsequent skills. In response to the trigger command of the intelligent rotation plan control, the intelligent rotation plan sub-interface is displayed, multi-dimensional input data of resident physicians is collected, and the rotation plan is generated and visualized by combining the skill tree, rotation cycle and teaching resource constraints, and interactive adjustment and hard constraint verification are supported; In response to the trigger command of the dynamic feedback control, a dynamic feedback sub-interface is displayed to collect feedback data in real time during the execution of the rotation scheme, calculate the comprehensive risk index, adjust the rotation scheme, and achieve closed-loop management.

2. The intelligent management and interactive method for resident physician rotation according to claim 1, characterized in that, The skill dependency graph management sub-interface includes a skill node addition control, a dependency relationship configuration control, and a skill tree display area; The skill dependency graph management sub-interface displays the skill dependency graph management sub-interface, constructs and maintains a skill dependency graph representing the logical dependencies between clinical operation skills, and generates and displays the resident physician's skill tree based on the skill dependency graph, including the following steps: In response to the trigger command for adding a control to the skill node, the skill node editing window is displayed, and the basic information of the new skill input by the user is received. The basic information includes the skill name, the professional base, the required study hours and the qualification standards. The new skill is then added as an independent node to the skill dependency graph. In response to the trigger command of the dependency configuration control, the dependency configuration window is displayed. Directed dependency edges are established between two skill nodes by dragging or selecting. During the establishment process, it is checked whether a loop is formed. If a loop is detected, saving is prohibited and the user is prompted to correct it, so as to ensure that the skill dependency graph always maintains a directed acyclic structure. Based on the skill dependency graph, a skill tree is generated for each resident physician, which includes automatically matching the corresponding prerequisite skills and subsequent skill paths according to the skill items that the resident physician has completed. The skill tree display area shows the skill tree of the selected resident physician.

3. The intelligent management and interactive method for resident physician rotation according to claim 2, characterized in that, The process of generating a skill tree for each resident physician based on the skill dependency graph includes the following steps: Obtain the set of skills that the current resident physician has completed; In the skill dependency graph, the nodes corresponding to the completed skill items are marked as mastered. Starting from all mastered nodes, traverse the skill dependency graph, identify all locked skill nodes whose prerequisite skills have been marked as mastered, and add the locked skill nodes to the unlockable skill set; Add the skill nodes from the unlockable skill set to the resident physician's skill tree and update the mastery status; Repeat the traversal and identification steps until the set of unlockable skills is empty, thereby generating a complete skill tree.

4. The intelligent management and interactive method for resident physician rotation according to claim 2, characterized in that, The skill dependency graph management sub-interface also includes version management controls; In response to the trigger command of the version management control, the version management window is displayed, which supports saving the current skill dependency graph version, publishing to the production environment, rolling back to the previous version, and branch management and independent maintenance of the skill dependency graph according to different professional bases.

5. The intelligent management and interactive method for resident physician rotation according to claim 1, characterized in that, The intelligent rotation plan sub-interface includes a data import control, an operation video upload control, a rotation plan generation control, and a plan adjustment control. The intelligent rotation plan sub-interface collects multi-dimensional input data from resident physicians, combines it with the skill tree, rotation cycle, and teaching resource constraints, generates and visualizes the rotation plan, and supports interactive adjustments and hard constraint verification, including the following steps: In response to the trigger command of the data import control, a data import window is displayed, and an electronic medical record system, simulation training platform or mobile scoring interface is configured and connected to obtain the resident physician's theoretical learning status, teaching score and personal basic information. In response to the trigger command of the operation video upload control, a video acquisition window is displayed to receive clinical operation videos uploaded by resident physicians or teaching staff, perform posture estimation and instrument trajectory recognition, and generate structured skill scores. In response to the trigger command of the rotation scheme generation control, based on the skill tree, the skill score, the upper limit of the rotation cycle, the department capacity limit and the status of teaching resources, the rule-based scheduling engine is invoked to generate a rotation scheme that meets the hard constraints of the prerequisite skill achievement, and the scheme is visualized in the form of a Gantt chart. In response to the trigger command of the scheme adjustment control, the user is allowed to drag and adjust the rotation period, and the scheme after adjustment is checked in real time to see if it violates hard constraints. When there is a conflict, a prompt message will pop up to ensure the compliance and executability of the rotation plan.

6. The intelligent management and interactive method for resident physician rotation according to claim 5, characterized in that, The process of invoking a rule-based scheduling engine to generate a round-robin scheme includes the following steps: Based on the skill tree and the skill scores, the set of rotation departments that each resident physician can currently enter is determined. A department is only included in the set of rotation departments that can be entered when the skill scores corresponding to all the prerequisite skills required by a department reach a preset threshold. Based on the preset rotation cycle and the maximum number of people that can be received in each department, the rotation time slots for resident physicians are allocated according to the first-come, first-served or preset rotation priority order. Synchronously match teaching resources to ensure that the assigned department has qualified teaching staff who are not currently overloaded with teaching duties during the corresponding time period. If, during the allocation process, there is insufficient department capacity, unavailable supervising teachers, or insufficient prior skills, the resident physician will be marked as waiting in line and given priority for re-arrangement in the next scheduling cycle.

7. The intelligent management and interactive method for resident physician rotation according to claim 1, characterized in that, The dynamic feedback sub-interface includes a workload feedback control; In the dynamic feedback sub-interface, feedback data during the execution of the rotation scheme is collected in real time, a comprehensive risk index is calculated, and the rotation scheme is adjusted, including the following steps: In response to the trigger command of the workload feedback control, a self-assessment form window is displayed to collect the workload and mental health data filled in by resident physicians. The workload includes the frequency of night shifts and the average number of surgeries participated in per day, and the mental health data includes the self-assessment results of mental health. Based on the workload and mental health data, the comprehensive risk index is calculated. When the comprehensive risk index exceeds a preset threshold, an early warning notification is automatically pushed to the management terminal, and corresponding intervention suggestions, the current skill tree, and historical rotation records are displayed in conjunction for teaching management personnel to make decisions.

8. The intelligent management and interactive method for resident physician rotation according to claim 7, characterized in that, The calculation of the comprehensive risk index based on the workload and mental health data specifically includes the following steps: Workload was quantitatively assessed by calculating the workload risk component based on the number of night shifts in the past two weeks, the frequency of departmental rotations per week, and the average number of clinical tasks per day. The mental health status was quantitatively assessed and mapped to a psychological risk component based on the standardized self-assessment scale scores submitted by resident physicians. The comprehensive risk index is obtained by summing the load risk component and the psychological risk component according to preset weights. When the comprehensive risk index exceeds the configured risk warning threshold, a tiered warning mechanism is triggered.

9. The intelligent management and interactive method for resident physician rotation according to claim 1, characterized in that, The main interface also includes a learning task distribution control; In response to the trigger command of the control for issuing the learning task, the learning task configuration window is displayed. When a certain rotation stage is activated, the associated teaching resource package is automatically retrieved from the hospital's digital education resource library according to the skill tag corresponding to the rotation stage. This package includes the corresponding textbook chapters, operation demonstration videos, clinical guidelines, case courseware, and accompanying online test questions. The teaching resource package is pushed to the simulation training platform of the corresponding resident physician, and a learning deadline is set; The resident physicians' learning completion status, video viewing time, and test scores are recorded as evidence of their theoretical learning.

10. A smart management and interactive system for resident physician rotation, characterized in that, Includes the following modules: The main interface module is configured to display the main interface for intelligent management of resident physician rotations; wherein, the main interface includes a skill dependency graph management control, an intelligent rotation plan control, and a dynamic feedback control; The skill dependency graph management module is configured to: in response to a trigger command on the skill dependency graph management control, display a skill dependency graph management sub-interface, construct and maintain a skill dependency graph representing the logical dependencies between clinical operation skills, and generate and display the resident physician's skill tree based on the skill dependency graph; The skill dependency graph is a directed acyclic graph, where nodes represent clinical operation skills and directed edges represent the dependency relationship between preceding and subsequent skills. The intelligent rotation plan module is configured to: in response to the trigger command of the intelligent rotation plan control, display the intelligent rotation plan sub-interface, collect multi-dimensional input data of resident physicians, and generate and visualize the rotation plan in combination with the skill tree, rotation cycle and teaching resource constraints, and support interactive adjustment and hard constraint verification; The dynamic feedback module is configured to: respond to the trigger command of the dynamic feedback control, display the dynamic feedback sub-interface, collect feedback data in real time during the execution of the rotation scheme, calculate the comprehensive risk index, adjust the rotation scheme, and realize closed-loop management.

Images