A training system for foreign surgical instrument nurses based on instrument atlas and deep learning.

By using a system based on an instrument atlas library and deep learning, the training priority of external surgical instruments is dynamically calculated, which solves the problem of uneven distribution of training resources in traditional systems and improves the efficiency of nurse training and emergency response capabilities.

CN123120232APending Publication Date: 2026-11-17WUXI PEOPLES HOSPITAL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611381791.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing training system for nurses using external surgical instruments cannot adjust the training priority according to the differentiated needs of different departments, resulting in insufficient training for high-frequency emergency instruments and excessive allocation of training resources for low-frequency instruments, which affects the efficiency of clinical emergency response.

Method used

A system based on an instrument atlas library and deep learning is adopted. Through nurse information identification, emergency department association analysis, scheduling linkage prediction and specialty instrument priority generation modules, the system dynamically calculates the training priority of instruments, generates a differentiated training list, and optimizes the allocation of training resources by combining graph neural networks and time series models.

Benefits of technology

This approach achieves a precise match between training content and actual clinical needs, improves nurses' mastery of high-frequency and critical instruments, enhances instrument identification accuracy and emergency response capabilities in multidisciplinary collaborative surgical scenarios, and optimizes the utilization efficiency of training resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120232A_ABST
    Figure CN123120232A_ABST
Patent Text Reader

Abstract

This invention discloses a training system for nurses using external surgical instruments based on an instrument atlas library and deep learning, belonging to the field of medical information technology. It aims to solve the technical problem of current training systems for nurses using external surgical instruments lacking the ability to adjust priorities for specific departments. The system includes an instrument atlas library module, a nurse information recognition module, an emergency department association analysis module, a scheduling linkage prediction module, and a specialty instrument priority generation module. Based on a time series prediction model, it outputs the types of surgeries nurses will participate in and cross-departmental collaboration information. By setting up the instrument atlas library module, nurse information recognition module, emergency department association analysis module, scheduling linkage prediction module, and specialty instrument priority generation module, this invention generates a differentiated training priority list, achieving precise matching of training content with actual clinical needs and effectively improving nurses' mastery of high-frequency and critical instruments.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of medical information technology, and more specifically, to a training system for nurses using foreign surgical instruments based on an instrument atlas library and deep learning. Background Technology

[0002] In modern hospital surgical management systems, there is a wide variety of external surgical instruments that are frequently updated. This is especially true in multidisciplinary surgeries and emergency surgeries, where nurses need to quickly master the structural characteristics, usage guidelines, and sterilization procedures for a large number of instruments from different departments. However, existing training methods mainly rely on static diagrams or manual instruction based on experience, which suffers from low learning efficiency, outdated knowledge, and the inability to quantify and assess individual differences.

[0003] Furthermore, the frequency of instrument use and the correlation of surgical procedures vary significantly across different departments. Traditional training systems cannot provide differentiated training arrangements based on nurses' departments, resulting in insufficient training for some high-frequency emergency instruments and excessive allocation of training resources for low-frequency instruments, further impacting clinical emergency response efficiency. Therefore, we propose a training system for nurses using external surgical instruments based on an instrument atlas library and deep learning. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art, adapt to practical needs, and provide a training system for nurses using external surgical instruments based on an instrument atlas library and deep learning, so as to solve the technical problem that the existing training system for nurses using external surgical instruments cannot adjust the training priority for specific departments.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a nurse training system for foreign surgical instruments based on an instrument atlas library and deep learning, including an instrument atlas library module, a nurse information recognition module, an emergency department association analysis module, a scheduling linkage prediction module, and a specialty instrument priority generation module; The nurse information identification module is used to obtain the nurse's personal information in the medical facility, including the department, job level, and historical surgical participation records. The emergency department association analysis module is used to identify whether the nurse's department belongs to an emergency surgery associated department and to construct a cross-departmental collaborative relationship map between emergency departments and other surgical departments; The scheduling linkage prediction module is used to acquire historical and future scheduling data of nurses, and output the types of surgeries that nurses will participate in in the future and cross-departmental collaboration information based on the time series prediction model. The specialized medical device priority generation module is connected to the nurse information identification module, the emergency department association analysis module, the scheduling linkage prediction module, and the medical device atlas library module. It is used to dynamically calculate the weight of external surgical instruments in the medical device atlas library based on the nurse's emergency department attributes, cross-departmental collaboration relationships, and predicted surgical procedure information, and generate a differentiated medical device training priority list.

[0006] Preferably, the emergency surgery-related departments include one or more of the following: emergency department, trauma surgery, neuroemergency department, thoracic surgery emergency department, and intensive care unit. The emergency department association analysis module constructs a relationship map between emergency departments and collaborating departments based on cross-departmental joint surgery data. The relationship map includes the frequency of joint surgeries, the frequency of instrument calls, and the frequency of surgical procedure sharing.

[0007] Preferably, the relationship graph is trained using a graph attention network model based on emergency weight constraints, and the embedded representation learning is performed on nurse nodes, emergency department nodes, collaborating department nodes, surgical procedure nodes, and instrument nodes. The graph attention network model introduces emergency association weight coefficients during node information propagation to enhance the edge weights corresponding to emergency department nodes, and constructs a dynamic weighted adjacency matrix based on the frequency of cross-department joint surgery, the frequency of equipment call-up, and the correlation of scheduling prediction. The system uses an emergency priority attention mechanism to weight and adjust the information propagation path between nodes to generate an equipment importance score. The system is based on an emergency weight constraint graph attention propagation algorithm.

[0008] Preferably, the graph attention propagation algorithm with emergency rescue weight constraints includes an emergency rescue weighted edge weight model, an emergency rescue priority attention mechanism, and an emergency rescue path priority propagation rule, wherein the emergency rescue weighted edge weight model operates based on the following formula:

[0009] in, For edge weights, , , These are the weighting coefficients, and , .

[0010] Preferably, the nurse shift scheduling sequence feature model includes historical shift data, current shift status, and future shift prediction data; The scheduling linkage prediction module uses a time series learning model based on long short-term memory network to encode the features of the nurses' cross-department scheduling sequence and generate cross-department surgical task association vectors. The cross-departmental surgical task association vector is then input into the specialty instrument priority generation module, serving as a time-dynamic weighting factor in the emergency weighted edge-weight model to adjust the instrument priority ranking results. This is generated based on an emergency-constrained cross-departmental sequence embedding algorithm.

[0011] in, This is a cross-departmental surgical task association vector. These are the latent state features of the shift scheduling sequence. This represents the cross-departmental correlation vector, where λ is the emergency reinforcement coefficient. This is the time weighting coefficient; The cross-departmental surgical task association vector is then input into the instrument priority generation module.

[0012] Preferably, the specialty device priority generation module calculates device priority based on a multi-factor weighted model, wherein the multi-factors include at least the emergency department association weight, the cross-departmental collaboration frequency weight, the scheduling prediction procedure weight, and the historical device error rate weight.

[0013] Preferably, the priority list of instruments is sorted according to the following rules: emergency high-frequency collaborative surgical instruments - instruments related to recent scheduling prediction - instruments for cross-departmental shared surgical procedures - instruments with high historical training error rates.

[0014] Preferably, the system further includes an intelligent instrument recognition module, which is used to extract features from instrument images based on a convolutional neural network model and output instrument category, structural features and model information.

[0015] Preferably, the convolutional neural network model includes one or more combined models of ResNet, YOLO object detection, and VisionTransformer.

[0016] Preferably, it also includes a data management module; The data management module is used to store, update, and manage versions of nurse training data, device atlas data, and the results of special device priority calculation. Based on the nurse's device identification error rate, training completion rate, and task execution feedback, the data management module dynamically adjusts the weight parameters of the special device priority generation module to achieve adaptive optimization of the training strategy.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention, through the establishment of an instrument atlas library module, a nurse information identification module, an emergency department association analysis module, a scheduling linkage prediction module, and a specialty instrument priority generation module, enables the system to dynamically calculate the weight of external surgical instruments based on nurses' department attributes, historical surgical records, and future scheduling prediction information. This generates a differentiated training priority list, achieving precise matching between training content and actual clinical needs, and effectively improving nurses' mastery efficiency of high-frequency and key instruments.

[0018] 2. This invention also constructs a relationship graph between emergency departments and collaborating departments, and uses graph neural networks for node embedding and edge weight propagation modeling. This effectively solves the problems of weak cross-departmental instrument association and inability to quantify collaborative relationships in traditional systems, thereby improving the recognition accuracy and training relevance of instruments in multi-departmental joint surgical scenarios.

[0019] 3. This invention also introduces a priority scheduling mechanism based on the nurse's department, which automatically assigns higher equipment training weight to nurses in emergency departments. This enables the system to dynamically adjust the allocation of training resources according to the clinical risk level, thereby solving the problem of insufficient emergency response capabilities caused by the average allocation in traditional training and significantly improving emergency response capabilities in critical care scenarios.

[0020] 4. This invention also uses a scheduling linkage prediction module to perform time series modeling and prediction of nurses' future surgical participation, enabling the system to identify in advance the types of surgical procedures and equipment requirements that nurses may encounter, thereby realizing the pre-emptive and dynamic optimization of training content, avoiding the problems of insufficient temporary training or redundant allocation of resources, and improving the overall efficiency of training resource utilization and clinical adaptability. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the training closed-loop process of the present invention. Detailed Implementation

[0022] like Figure 1 As shown, the present invention relates to a nurse training system for foreign surgical instruments based on an instrument atlas library and deep learning, including an instrument atlas library module, a nurse information identification module, an emergency department association analysis module, a scheduling linkage prediction module, and a specialty instrument priority generation module. The nurse information identification module is used to obtain nurses' personal information in medical settings, including their department, job level, and historical surgical participation records. The emergency department association analysis module is used to identify whether the nurse's department belongs to an emergency surgery-related department and to construct a cross-departmental collaboration relationship map between emergency departments and other surgical departments; The scheduling linkage prediction module is used to obtain historical and future scheduling data of nurses, and outputs the types of surgeries that nurses will participate in in the future and cross-departmental collaboration information based on the time series prediction model. The specialty medical device priority generation module is connected to the nurse information identification module, the emergency department association analysis module, the scheduling linkage prediction module, and the medical device atlas library module. It is used to dynamically calculate the weight of foreign surgical instruments in the medical device atlas library based on the nurse's emergency department attributes, cross-departmental collaboration relationships, and predicted surgical procedure information, and generate a differentiated medical device training priority list.

[0023] Specifically, the departments associated with emergency surgery include one or more of the following: emergency department, trauma surgery, neuroemergency department, thoracic surgery, and intensive care unit. The emergency department association analysis module constructs a relationship map between emergency departments and collaborating departments based on cross-departmental joint surgery data. The relationship map includes the frequency of joint surgeries, the frequency of instrument calls, and the frequency of surgical procedure sharing.

[0024] More specifically, the relationship graph is trained using a graph attention network model based on emergency weight constraints, and embedding representation learning is performed on nurse nodes, emergency department nodes, collaborating department nodes, surgical procedure nodes, and instrument nodes. Among them, the graph attention network model introduces emergency association weight coefficients in the process of node information propagation to enhance the edge weights corresponding to emergency department nodes, and constructs a dynamic weighted adjacency matrix based on the frequency of cross-department joint surgery, the frequency of equipment call and the correlation of scheduling prediction. The system uses an emergency priority attention mechanism to weight and adjust the information propagation path between nodes to generate an equipment importance score. The system is based on an emergency weight constraint graph attention propagation algorithm.

[0025] Furthermore, the graph attention propagation algorithm with emergency response weight constraints includes an emergency response weighted edge weight model, an emergency response priority attention mechanism, and an emergency response path priority propagation rule. The emergency rescue weighted edge-weight model operates based on the following formula:

[0026] in, For edge weights, , , These are the weighting coefficients, and , ; The emergency priority attention mechanism operates based on the following formula:

[0027] in, For emergency response intensity, This refers to the emergency response enhancement coefficient. The emergency route priority propagation rule operates based on the following formula.

[0028] in, As a critical point for emergency response, for Layer nodes Feature representation, For nodes The set of adjacent nodes.

[0029] Furthermore, the nurse shift scheduling sequence feature model includes historical shift data, current shift status, and future shift prediction data; Among them, the scheduling linkage prediction module adopts a time series learning model based on long short-term memory network to encode the features of the nurses' cross-department scheduling sequence and generate cross-department surgical task association vectors; The cross-departmental surgical task association vector is input into the specialty instrument priority generation module, which serves as a time-dynamic weighting factor in the emergency weighted edge weight model to adjust the instrument priority ranking results. This is generated based on a cross-departmental sequence embedding algorithm for emergency constraints.

[0030] in, This is a cross-departmental surgical task association vector. These are the latent state features of the shift scheduling sequence. This represents the cross-departmental correlation vector, where λ is the emergency reinforcement coefficient. This is the time weighting coefficient; The cross-departmental surgical task association vector is then input into the instrument priority generation module.

[0031] It is worth noting that the specialty device priority generation module calculates device priority based on a multi-factor weighted model. The multi-factors include at least the emergency department association weight, the cross-departmental collaboration frequency weight, the scheduling prediction procedure weight, and the historical device error rate weight. The priority list of instruments is sorted according to the following rules: emergency high-frequency collaborative surgical instruments - instruments related to recent scheduling predictions - instruments for cross-departmental shared surgical procedures - instruments with high historical training error rates.

[0032] It is worth mentioning that the system also includes an intelligent medical device recognition module, which is used to extract features from medical device images based on a convolutional neural network model and output medical device category, structural features and model information; Convolutional neural network models include one or more combinations of ResNet, YOLO object detection, and VisionTransformer models.

[0033] It is worth noting that it also includes a data management module, which is used to store, update and manage versions of nurse training data, equipment atlas data and priority calculation results of specialty equipment; The data management module dynamically adjusts the weight parameters of the specialty device priority generation module based on nurses' device identification error rate, training completion rate, and task execution feedback, so as to achieve adaptive optimization of the training strategy.

[0034] It should be noted that the following terms in this specification have the following meanings: External surgical instruments: refers to reusable surgical instruments that are not routinely provided by our hospital, are supplied by external suppliers, and are used for specific specialty or emergency surgeries; Instrument Atlas Library: A database module that stores images, structures, specifications, usage procedures, and inventory standards for imported surgical instruments; Cross-departmental collaboration relationship graph: a relationship network constructed with departments, nurses, procedures, and instruments as nodes and collaboration intensity and sharing frequency as edge weights; Dynamic weight calculation: A method for calculating the importance weight of external surgical instrument training in real time based on nurses' departmental attributes, shift forecasts, and cross-departmental collaboration relationships; Differentiated training priority list: Automatically generated equipment training lists with different learning sequences and key contents, tailored to the different positions and scenarios of nurses; Edge weight propagation mechanism: In graph neural networks, weights are propagated and weighted along the edges of the relation graph to obtain the calculation rules for the importance score of the device.

[0035] Example 1: This example provides a training system for external surgical instruments for emergency department nurses. The module operation logic is as follows: Nurse Information Identification Module: Collects the target nurse's personal information, which is from the Emergency Department, is at the N2 level, and has participated in a total of 86 emergency surgeries and trauma emergency surgeries in the past.

[0036] Emergency Department Association Analysis Module: Determines that the nurse belongs to an emergency surgery related department; Based on cross-departmental joint surgery data from the past year, constructs a collaborative relationship map between the emergency department and trauma surgery, neurosurgery, and thoracic surgery, with edge weights being the frequency of joint surgeries, the frequency of instrument sharing, and the degree of surgical procedure matching.

[0037] The scheduling linkage prediction module imports the nurse's schedule for the past 3 months and uses a time series model to predict the three main types of surgeries she will participate in for the next 2 weeks: emergency debridement, trauma fracture fixation, and emergency thoracotomy.

[0038] Specialty equipment priority generation module: Combining emergency attributes, cross-departmental collaboration relationships and predicted surgical procedures, it performs dynamic weight calculations on foreign equipment in the equipment atlas library to generate a priority Top 10 training list: trauma fixation equipment, emergency thoracic equipment, neurological emergency equipment, rapid disinfection equipment, emergency hemostasis equipment, etc.

[0039] Medical Device Atlas Library Module: Pushes high-definition atlases, structural descriptions, inventory procedures, usage contraindications, and packaging specifications of corresponding medical devices according to priority.

[0040] Technical benefits: The system automatically matches high-frequency emergency external instruments to emergency department nurses, skipping low-frequency general surgical instruments, shortening training time, improving instrument recognition accuracy, and meeting the needs of rapid emergency response.

[0041] Example 2: This example provides a training system for external surgical instruments for nurses in the intensive care unit. The module operation logic is as follows: Nurse Information Identification Module: Collects nurse information, whose department is the Intensive Care Unit, whose job level is N3, and who has participated in 42 joint rescue surgeries involving multiple departments in the past.

[0042] Emergency Department Association Analysis Module: Determines the ICU as an emergency-related department; trains a cross-departmental collaborative relationship graph using a graph neural network model, performs vector embedding on nurse nodes, department nodes, procedure nodes, and instrument nodes, and obtains the instrument importance score by propagating according to edge weights.

[0043] The scheduling linkage prediction module analyzes historical and future schedules to predict whether the nurse will participate in two interdepartmental joint surgeries next week: one between the ICU and the general surgery department, and the other between the ICU and the neurosurgery department.

[0044] Specialty equipment priority generation module: integrates departmental attributes, collaborative relationships, and predicted surgical procedures to dynamically calculate the weight of external equipment and prioritize the delivery of external equipment for intensive care, neurosurgical emergency equipment, and general surgery combined surgical equipment.

[0045] Medical Instrument Atlas Module: Provides 3D atlases of corresponding medical instruments, inventory guides, cross-departmental handover procedures, and emergency troubleshooting.

[0046] Technical benefits: The system accurately matches cross-departmental collaborative scenarios, automatically updates training content, ensures zero errors in the handover of external instruments by nurses, shortens surgical preparation time, and reduces the risk of instrument mismatch.

[0047] The embodiments disclosed in this invention are preferred embodiments, but are not limited thereto. Those skilled in the art can easily understand the spirit of this invention based on the above embodiments and make different extensions and variations, but as long as they do not depart from the spirit of this invention, they are all within the protection scope of this invention.

Claims

1. A training system for nurses using imported surgical instruments based on an instrument atlas library and deep learning, characterized in that, It includes a medical device atlas library module, a nurse information identification module, an emergency department association analysis module, a scheduling linkage prediction module, and a specialty medical device priority generation module; The nurse information identification module is used to obtain the nurse's personal information in the medical facility, including the department, job level, and historical surgical participation records. The emergency department association analysis module is used to identify whether the nurse's department belongs to an emergency surgery associated department and to construct a cross-departmental collaborative relationship map between emergency departments and other surgical departments; The scheduling linkage prediction module is used to acquire historical and future scheduling data of nurses, and output the types of surgeries that nurses will participate in in the future and cross-departmental collaboration information based on the time series prediction model. The specialized medical device priority generation module is used to dynamically calculate the weights of foreign surgical instruments in the medical device atlas library based on the nurse's emergency department attributes, cross-departmental collaboration relationships, and predicted surgical procedure information, and to generate a differentiated medical device training priority list. 2.The foreign surgical instrument nurse training system based on an instrument atlas library and deep learning according to claim 1, wherein The departments associated with the emergency surgery include one or more of the following: emergency department, trauma surgery, neuroemergency department, thoracic surgery, and intensive care unit. The emergency department association analysis module constructs a relationship map between emergency departments and collaborating departments based on cross-departmental joint surgery data. The relationship map includes the frequency of joint surgeries, the frequency of instrument calls, and the frequency of surgical procedure sharing. 3.The foreign surgical instrument nurse training system based on an instrument atlas library and deep learning according to claim 2, characterized in that, The relationship graph is trained using a graph attention network model based on emergency weight constraints, and embedding representation learning is performed on nurse nodes, emergency department nodes, collaborating department nodes, surgical procedure nodes, and instrument nodes. The graph attention network model introduces emergency association weight coefficients during node information propagation to enhance the edge weights corresponding to emergency department nodes, and constructs a dynamic weighted adjacency matrix based on the frequency of cross-department joint surgery, the frequency of equipment call-up, and the correlation of scheduling prediction. The system uses an emergency priority attention mechanism to weight and adjust the information propagation path between nodes to generate an equipment importance score. The system is based on an emergency weight constraint graph attention propagation algorithm.

4. The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 3, characterized in that, The graph attention propagation algorithm with emergency rescue weight constraints includes an emergency rescue weighted edge weight model, an emergency rescue priority attention mechanism, and an emergency rescue path priority propagation rule. The emergency rescue weighted edge weight model operates based on the following formula: ; in, For edge weights, , , These are the weighting coefficients, and , . 5.The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 1, wherein, The nurse shift scheduling sequence feature model includes historical shift data, current shift status, and future shift prediction data; The scheduling linkage prediction module uses a time series learning model based on long short-term memory network to encode the features of the nurses' cross-department scheduling sequence and generate cross-department surgical task association vectors. The cross-departmental surgical task association vector is then input into the specialty instrument priority generation module, serving as a time-dynamic weighting factor in the emergency weighted edge-weight model to adjust the instrument priority ranking results. This is generated based on an emergency-constrained cross-departmental sequence embedding algorithm. ; in, This is a cross-departmental surgical task association vector. These are the latent state features of the shift scheduling sequence. This represents the cross-departmental correlation vector, where λ is the emergency reinforcement coefficient. This is the time weighting coefficient; The cross-departmental surgical task association vector is then input into the instrument priority generation module.

6. The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 5, characterized in that, The specialized medical device priority generation module calculates the priority of medical devices based on a multi-factor weighted model. The multi-factors include at least the emergency department association weight, the cross-departmental collaboration frequency weight, the scheduling prediction procedure weight, and the historical medical device error rate weight. 7.The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 6, wherein, The priority list of instruments is sorted according to the following rules: emergency high-frequency collaborative surgical instruments - instruments related to recent scheduling predictions - instruments for cross-departmental shared surgical procedures - instruments with high historical training error rates. 8.The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 1, wherein, The system also includes an intelligent instrument recognition module, which is used to extract features from instrument images based on a convolutional neural network model and output instrument category, structural features and model information. 9.The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 8, wherein, The convolutional neural network model includes one or more combinations of ResNet, YOLO object detection, and VisionTransformer models. 10.The foreign surgical instrument nurse training system based on instrument atlas library and deep learning according to claim 1, wherein, It also includes a data management module; The data management module is used to store, update, and manage versions of nurse training data, equipment atlas data, and priority calculation results for specialty equipment. The data management module dynamically adjusts the weight parameters of the specialty instrument priority generation module based on the nurse's instrument identification error rate, training completion rate, and task execution feedback, so as to achieve adaptive optimization of the training strategy.