Operation management method and system for operating room data analysis and visualization

Abstract

The invention relates to the technical field of medical informatization management, and discloses an operating room data analysis and visualization operation management method and system, and the method comprises the steps: collecting the related data of an operating room, carrying out the data preprocessing, and obtaining the actual load of the current operating room according to the related data of the operating room, carrying out operation type matching on the operation types in the standard operation type library and the standard operation type library; an operating room load prediction model is constructed, operating room loads are dynamically predicted by adjusting parameters of the operating room load prediction model so as to evaluate operating room resources, and nurse resource evaluation and department resource evaluation are performed according to the operating room related data; the operating room resource evaluation result, the nurse resource evaluation result and the department resource evaluation result are integrated and visually displayed, visual analysis and real-time interaction of resource load and configuration condition evaluation are provided, scientificity and real-time performance of operating room resource scheduling are improved, nurse scheduling and department resource configuration are optimized, and the operating room resource scheduling efficiency is improved. And resource idleness and overload risks are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of medical informatization management, and particularly to an operation management method and system for operating room data analysis and visualization. Background Art

[0002] With the acceleration of the medical informatization process, the intelligent demand for the hospital operating room operation management system has become increasingly prominent.

[0003] However, there are still significant defects in the existing technologies in practical applications, which are specifically manifested in the following three aspects: (1) Insufficient data quality: Outliers of operation duration are not effectively cleaned, it is difficult to split the combined operation methods due to non-standard manual input, the matching efficiency of operation method names with the standard library (such as ICD9-CM3) is low, and conflicts are caused by duplicate employee numbers and names in historical data; (2) Single resource evaluation method: The evaluation of nurses' capabilities in traditional systems relies on manual experience and lacks quantitative indicators; The analysis of departmental resource distribution is limited to simple statistics, making it difficult to identify doctors' expertise and resource bottlenecks; The prediction of operating room load lacks a dynamic model, resulting in resource idleness or overload; (3) Weak visualization ability: Existing tools mostly use static charts, which cannot support multi-dimensional data interaction, and the rendering efficiency of large-scale data is low, making it difficult to display complex indicators in real time.

[0004] The above problems lead to low utilization efficiency of hospital operating room resources and lagging management decisions. Therefore, there is an urgent need for a comprehensive management platform integrating data cleaning, intelligent evaluation, and dynamic visualization to improve the scientificity and real-time nature of resource allocation. Summary of the Invention

[0005] The purpose of the present invention is to solve the above-mentioned disadvantages existing in the prior art, and to provide an operation management method and system for operating room data analysis and visualization, which cover medical data preprocessing, multi-dimensional resource evaluation, and interactive visualization, aiming to achieve efficient scheduling and intelligent management of hospital operating room resources.

[0006] On the one hand, an operation management method for operating room data analysis and visualization is provided, including the following steps: S1: Collect relevant data of the operating room and perform data preprocessing, obtain the actual load of the current operating room according to the relevant data of the operating room, convert the operation method name therein into a semantic vector through an Embedding model, and realize operation method matching with the standardized operation method library in combination with the cosine similarity; S2: Based on the actual load of the operating room, the trend factor, and the seasonal adjustment factor, construct an operating room load prediction model. Dynamically predict the operating room load in the next time period by adjusting the parameters of the operating room load prediction model. Evaluate the operating room resources according to the operating room load prediction results, and then conduct nurse resource evaluation and department resource evaluation based on the relevant operating room data; S3: Integrate the evaluation results of the operating room resources, nurse resources, and department resources and conduct visual display to provide an intuitive analysis and real-time interaction for resource load and configuration assessment.

[0007] Further, in step S1, the collection of relevant operating room data and data preprocessing further includes: Convert the time data in the relevant operating room data into the datetime format for calculating time differences, obtain the duration of each operation, the number of days the operating room is open, and the total working hours of the operating room, and calculate the actual load of the operating room therefrom; Group the surgical procedures in the relevant operating room data according to the procedure codes, calculate the average and standard deviation of the operation duration, the table duration, and the difference for each procedure, and obtain the upper and lower limits of the operation duration, the table duration, and the difference duration for each procedure according to the 6 Sigma principle. Use these limits to screen out abnormal records that do not meet the standards; Establish an anti-duplicate name conflict mechanism, and use the combination of name and job number to ensure the unique identification of each nurse and doctor in the relevant operating room data.

[0008] Preferably, in step S1, the matching of the surgical procedure with the standardized procedure library further includes: When entering the surgical procedure, split the name of the composite surgical procedure through regular expressions; Load the procedure data in the ICD9-CM3 standard library, including the name of each procedure, and use the procedure data as input for pre-training of the Embedding model; Use the pre-trained Embedding model to convert the name of each procedure in the ICD9-CM3 standard library into an embedding vector to generate a standard procedure embedding vector set. At the same time, use this pre-trained model to also convert the query procedure input by the user into an embedding vector to obtain a query embedding vector; Calculate the cosine similarity between the query embedding vector and all the embedding vectors in the standard procedure embedding vector set, and evaluate the similarity between the query procedure and each standard procedure; Return the standard procedure most similar to the query procedure according to the calculated similarity score.

[0009] Further, in step S2, constructing an operating room load prediction model based on the actual load, trend factor, and seasonal adjustment factor of the operating room further includes: S21: Obtain the operating room load data for the historical period and ensure consistent time granularity. Decompose the data into a trend term, seasonal term, and residual term through the STL decomposition method, and calculate the trend factor using the linear regression method , and obtain the seasonal adjustment factor by quantifying the periodic influence through the seasonal index ; S22: Construct an operating room load prediction model, and the expression of the model formula is as follows: , where, represents the operating room load for the next period, represents the operating room load for the current period, , , are all empirical parameters that can be adjusted according to historical data, and set the initial weights of the parameters , , ; S23: Train the model, use the mean squared error MSE to measure the deviation between the predicted value and the actual value, and use the gradient descent method to adjust the parameters , , , and minimize the MSE to optimize the operating room load prediction model.

[0010] Preferably, in step S2, dynamically predicting the operating room load for the next period by adjusting the parameters of the operating room load prediction model further includes: Update the result influence factor through a sliding window. After adding the actual operating room load data for each new period, recalculate the trend factor and seasonal factor; Perform online learning of the model using incremental training, and fine-tune the parameters of the model through exponential weighted adjustment based on the new data; Substitute the latest parameters and factors into the operating room load prediction model to predict the operating room load for the next period.

[0011] Further, in step S2, the nurse resource assessment includes nurse ability assessment and nurse work efficiency assessment, and further includes: The assessment of the nurse's ability includes the ability assessment of the number of surgeries participated by the nurse, department coverage, and the participation rate in high-level surgeries. The nurse's ability level is rated by calculating the nurse ability index , and the calculated ability index is converted into the star rating of the nurse according to a preset standard, and the calculation formula is as follows: , wherein, , , is the weight coefficient of the corresponding evaluation dimension, which is used to dynamically set and adjust the contribution of different dimensions to the ability index. is the standardized number of surgeries, is the standardized department coverage, is the participation rate in high - level surgeries; Determine the evaluation indicators of the nurses' work efficiency, including the total follow - up duration , the total number of follow - up times , the average working duration per surgery , the number of follow - up days , the average follow - up duration per day and the average number of follow - up times per day . Standardize each indicator and calculate the comprehensive work efficiency score of each nurse . The calculation formula is as follows: , wherein, , , , , , is the weight of each indicator calculated by the Analytic Hierarchy Process (AHP).

[0012] Furthermore, in step S2, the department resource evaluation includes the analysis of the distribution of surgical procedures and surgical grades and the cross - analysis of the abilities of the chief surgeons and surgical procedures. Among them, the analysis of the distribution of surgical procedures and surgical grades specifically includes: Collect the historical data of the operating room, including the operation date, operation grade, name of surgical procedure, department name, chief surgeon, and participating nurses. Group the operation data of each department according to the operation grade, dynamically detect the operation grade field, and count the number of surgeries in different operation grades for each department; Use a radar chart to display the performance of each department in surgeries of different grades and the strengths of each department in surgeries of different grades, and then stack the number of surgeries of each department according to the operation grade to display the load of each department in different operation grades; The cross - analysis of the abilities of the chief surgeons and surgical procedures specifically includes: Extract information including the names of the doctors participating in the surgery, the types of surgical procedures participated by the doctors, and the departments of the surgeries participated by the doctors from the input surgical - related data; By creating a cross - data structure between doctors and surgical procedures, record the types of surgeries participated by each doctor, and match the surgical procedures that each doctor is good at with the requirements of the operating room.

[0013] Further, in step S3, integrating and visually displaying the evaluation results of the operating room resources, nurse resources, and department resources further includes: The nurse ergonomics score is a specific value, which is visualized through a radar chart and a violin chart to display the distribution of each ergonomics index and the comprehensive performance of each nurse; The analysis of the distribution of surgical procedures and surgical grades generates a nested data structure, which is adapted to a radar chart and a stacked bar chart. The radar chart is used to display the performance of each department in surgeries of different grades and the strengths of each department in surgeries of different grades, and the stacked bar chart is used to intuitively display the load of each department at different surgical grades; The cross-analysis result of the attending surgeon and surgical procedure capabilities is displayed through a cross Venn diagram to show the work distribution of doctors in specific surgical procedures and departments, helping the hospital identify surgical expertise and resource allocation bottlenecks, and optimizing doctor scheduling and resource allocation; In the evaluation of the operating room resource utilization efficiency, the working hours, opening days, and load rate information of the operating room are displayed through charts to help managers monitor the resource usage in real time. The average working hours of each operating room are displayed through a bar chart of the average working hours of each operating room, the opening days of each operating room are displayed through a bar chart of the opening days of each operating room within a certain period, the number of working times of each doctor in each operating room is displayed through a bar chart of the number of doctor working times, providing a basis for doctor scheduling and resource optimization. The predicted operating room load curve is displayed through a prediction curve and compared with the actual load to help predict future demands.

[0014] On the other hand, an operation management system for operating room data analysis and visualization is provided, including: A data preprocessing module, configured to collect operating room-related data and perform data preprocessing, obtain the actual load of the current operating room according to the operating room-related data, convert the surgical procedure name therein into a semantic vector through an Embedding model, and perform surgical procedure matching with a standardized surgical procedure library in combination with cosine similarity; A multi-dimensional resource evaluation module, configured to construct an operating room load prediction model based on the actual load of the operating room, a trend factor, and a seasonal adjustment factor, dynamically predict the operating room load in the next time period by adjusting the parameters of the operating room load prediction model, evaluate the operating room resources according to the operating room load prediction result, and then evaluate the nurse resources and department resources according to the operating room-related data; An interactive visualization module, configured to integrate and visually display the evaluation results of the operating room resources, nurse resources, and department resources, and provide an intuitive analysis and real-time interaction for the evaluation of resource load and configuration.

[0015] In addition, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, it implements the operation management method for operating room data analysis and visualization described in any one of the above.

[0016] Meanwhile, an electronic device is provided, including: one or more processors; a storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the operation management method for operating room data analysis and visualization described in any one of the above.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention cleans abnormal duration data through the 6 Sigma principle, splits compound surgical procedures using regular expressions, combines models to achieve semantic matching between surgical procedures and the ICD9-CM3 standard library, and uses unique identifiers to solve nurse data conflicts; The present invention constructs an operating room load prediction model, dynamically predicts the operating room load in the next time period by adjusting the parameters of the operating room load prediction model, and realizes dynamic prediction and evaluation of the operating room load based on historical data and trend factors; The present invention constructs a nurse ability index model based on the number of surgeries, department coverage, and high-level surgery participation rate, and an ergonomic comprehensive scoring model, and combines Min-Max standardization and the analytic hierarchy process to support the analysis of departmental surgery distribution; The present invention realizes real-time dynamic display of multi-dimensional data through an interactive visualization engine module, which integrates interactive charts such as radar charts, stacked bar charts, cross Venn diagrams, and violin diagrams. Description of the Drawings

[0018] The drawings are used to provide further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 It is a flowchart of an operation management method for operating room data analysis and visualization according to the present invention; Figure 2 It is a structural block diagram of an operation management system for operating room data analysis and visualization according to the present invention. Detailed Embodiments

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

[0020] The core technologies of this invention cover medical data preprocessing (including anomaly cleaning, compound surgical procedure splitting, and surgical procedure standardization matching), multi-dimensional resource evaluation (nurse ability and work efficiency quantification, department surgical distribution analysis, operating room load prediction), and interactive visualization, aiming to improve the scientificity and real-time nature of operating room resource scheduling, optimize nurse scheduling and department resource allocation, and reduce the risks of resource idleness and overload.

[0021] The following will illustrate the specific implementation manners of this invention with reference to the accompanying drawings and embodiments.

[0022] First Embodiment Please refer to Figure 1 , an operation management method for operating room data analysis and visualization provided in this embodiment. The technical solution includes the following steps: S1: Collect relevant operating room data and perform data preprocessing. Obtain the actual load of the current operating room based on the relevant operating room data, convert the surgical procedure names therein into semantic vectors through an Embedding model, and perform surgical procedure matching with a standardized surgical procedure library in combination with cosine similarity; S2: Construct an operating room load prediction model based on the actual operating room load, trend factor, and seasonal adjustment factor. Dynamically predict the operating room load for the next time period by adjusting the parameters of the operating room load prediction model. Evaluate the operating room resources based on the operating room load prediction result, and then evaluate the nurse resources and department resources according to the relevant operating room data; S3: Integrate the evaluation results of the operating room resources, nurse resources, and department resources and perform visual display to provide an intuitive analysis and real-time interaction for resource load and configuration evaluation.

[0023] Among them, in step S1, the collecting relevant operating room data and performing data preprocessing further includes: Convert the time data in the relevant operating room data into the datetime format for calculating time differences, obtain the duration of each operation, the number of days the operating room is open, and the total working hours of the operating room, and calculate the actual load of the operating room therefrom; Group the surgical procedures in the operation room - related data according to the procedure code (ICD9), and calculate the average and standard deviation of the operation duration, case duration, and difference for each procedure. Obtain the upper and lower bounds of the operation duration, case duration, and difference duration for each procedure according to the 6 Sigma principle, and use these bounds to screen out abnormal records that do not meet the standards; Establish a mechanism to prevent duplicate names and conflicts, and use a combination of name and staff number to ensure the unique identification of each nurse and doctor in the operation room - related data.

[0024] In this embodiment, the detection and cleaning of abnormal operation durations are important steps in data pre - processing, aiming to remove invalid data and ensure data quality. We use the 6 Sigma principle to detect and clean abnormal values, which is a commonly used statistical method, usually used to identify and process data that deviates from the normal range.

[0025] When entering the surgical procedure in step S1, a combined surgical procedure refers to performing multiple different surgical operations in one operation, and these surgical operations may be composed of different procedures. During the input process of surgical data, due to the irregular manual filling of the separator, the extraction of combined procedure information may be inaccurate. Therefore, we split the combined surgical procedures through regular expressions. In this embodiment, the "+" is used as the separator for different procedures in the combined procedure data, and we can split these data through regular expressions; Establish a mechanism to prevent duplicate names and conflicts, and use a combination of name and staff number to ensure the unique identification of each nurse and doctor. In the hospital information system (HIS), there may be duplicate problems with staff numbers and names in historical surgical data, especially during data migration or manual input. To solve this problem, we need to establish a mechanism to prevent duplicate names and conflicts to ensure the unique identification of each nurse. We use a combination of "name, staff number" to solve the problem of duplicate names.

[0026] Next, we match the entered surgical procedure with the standardized procedure library, including: When entering the surgical procedure, split the name of the combined surgical procedure through regular expressions; Load the procedure data in the ICD9 - CM3 standard library, including the name of each procedure, and use the procedure data as input for pre - training of the Embedding model; Use the pre - trained Embedding model to convert the name of each procedure in the ICD9 - CM3 standard library into an embedding vector to generate a set of standard procedure embedding vectors. At the same time, use this pre - trained model to also convert the query procedure entered by the user into an embedding vector to obtain a query embedding vector; Calculate the cosine similarity between the query embedding vector and all the embedding vectors in the set of standard surgical procedure embedding vectors, and evaluate the similarity between the query surgical procedure and each standard surgical procedure; Return the standard surgical procedure that is most similar to the query surgical procedure according to the calculated similarity score.

[0027] Specifically, in this embodiment, we use an Embedding model to implement surgical procedure matching. In surgical data processing, surgical procedure matching is an important step, which can compare the surgical procedures recorded in the hospital with a standardized surgical procedure library (such as the ICD9-CM3 standard library) to ensure the consistency and accuracy of the data. To improve the efficiency and accuracy of surgical procedure matching, we adopt an embedding model, which converts the surgical procedure name into a vector representation, and then methods such as cosine similarity can be used for comparison. The steps are briefly described as follows: (1) Load surgical procedure data: Load the surgical procedure data in the ICD9-CM3 standard library, including the name of each surgical procedure (such as the fopname column). These surgical procedure names will be used as inputs for the model to process; (2) Convert the surgical procedure name into an embedding vector: We use a pre-trained sentence embedding model (such as SentenceTransformer) to convert each surgical procedure name into an embedding vector. The embedding vector is a high-dimensional floating-point array, which can capture the semantic information of the surgical procedure name and ensure that the vectors of similar surgical procedure names are close in space. For example, after converting "heart surgery" and "coronary artery bypass surgery" into vectors using the SentenceTransformer model, their embedding vectors should be similar in the vector space; (3) Calculate the similarity between the query and the surgical procedure library: The user inputs a query surgical procedure (such as "eyelid suture"), and we also use the model to convert it into an embedding vector. Next, by calculating the cosine similarity between the query embedding vector and all the surgical procedure embedding vectors in the ICD9-CM3 standard library, we can evaluate the similarity between the query surgical procedure and each standard surgical procedure; (4) Return the most similar surgical procedure: According to the calculated similarity score, we can return the standard surgical procedure that is most similar to the query surgical procedure. This process can help us automatically compare the surgical procedures entered in the hospital with the standard surgical procedure library to ensure the consistency of the surgical procedure data.

[0028] Then, further including constructing an operating room load prediction model based on the actual load of the operating room, the trend factor, and the seasonal adjustment factor: S21: Obtain the operating room load data for the historical period and ensure that the time granularity is consistent. Decompose the data into a trend term, a seasonal term, and a residual term by the STL decomposition method, and calculate the trend factor using the linear regression method , the trend factor is the load growth rate calculated based on historical data (e.g., the load growth in the past few months), and the calculation formula is as follows: , If there is an obvious upward trend in the load of the operating room, this factor will be positive; otherwise, it will be negative. And the seasonal adjustment factor is obtained by quantifying the periodic influence through the seasonal index , according to the influence of factors such as seasonal changes and holidays on the load (e.g., the load adjustment during the Spring Festival holiday); S22: Construct an operating room load prediction model, and the model formula expression is as follows: , wherein, represents the load of the operating room in the next period, represents the load of the operating room in the current period, , , are all empirical parameters, which can be adjusted according to historical data, and set the parameters , , as the initial weights; S23: Train the model, use the mean square error MSE to measure the deviation between the predicted value and the actual value, and use the gradient descent method to adjust the parameters , , , and minimize MSE to optimize the operating room load prediction model.

[0029] Among them, the actual load L of the operating room is calculated according to the number of days D the operating room is open, the total working hours TH of the operating room, and the actual opening hours AH, and the calculation formula is as follows: , Among them, the total working hours TH of the operating room is obtained by calculating the sum of the durations of each operation in the operating room, the actual opening hours AH are obtained according to different operation arrangements in the hospital, and the number of days D the operating room is open is obtained by checking the usage dates of each operating room.

[0030] By counting the number of times doctors work in different operating rooms, it provides a basis for optimizing doctor scheduling. According to the work records of doctors, the number of times of work in each operating room is generated: the number of times each doctor works in each operating room is calculated by aggregating operation data to ensure an accurate reflection of the work distribution of doctors.

[0031] Specifically, the operation working hours are calculated by the entry time and exit time of the operation to obtain the working hours of each operation, and then converted into hours; the number of days the operating room is open is counted by checking the usage dates of each operating room; the total working hours of the operating room is the sum of the working hours of each operation in the operating rooms; the actual load of the operating room: that is, the ratio of the total working hours to the total available hours of the operating room; the actual opening hours can be obtained according to different operation arrangements in the hospital. For example, the opening hours of the operating room are sometimes 12 hours, 16 hours, or 24 hours a day.

[0032] On this basis, dynamically predicting the operating room load in the next time period by adjusting the parameters of the operating room load prediction model further includes: Updating the result influence factor through a sliding window. After adding the actual load data of the operating room for each new time period, recalculate the trend factor and seasonal factor; Adopt incremental training for online learning of the model, and fine-tune the parameters of the model through exponential weighted adjustment based on new data; Substitute the latest parameters and factors into the operating room load prediction model to predict the operating room load in the next time period.

[0033] Furthermore, in step S2, the nurse resource assessment includes nurse ability assessment and nurse work efficiency assessment, and further includes: The assessment of the nurse's ability includes the ability assessment of the number of operations the nurse participates in, department coverage, and the participation rate in high-level operations. By calculating the nurse ability index Assess the ability level of the nurse, and the calculated ability index Is converted into the star rating of the nurse according to a preset standard. The calculation formula is as follows: , where , , Is the weight coefficient of the corresponding evaluation dimension, used to dynamically set and adjust the contribution of different dimensions to the ability index, Is the standardized number of operations, Is the standardized department coverage, Is the participation rate in high-level operations, , , The calculation formula of where Is the total number of operations the nurse participates in, Is the number of operations of the nurse with the most operations in the hospital, Is the number of departments the nurse participates in, is the total number of departments in the hospital, is the number of high - level surgeries participated by nurses, is the total number of surgeries participated by nurses; Determine the evaluation indicators for the work efficiency of the nurses, including the total follow - up duration , the total number of follow - up times , the average working duration per surgery , the number of follow - up days , the average daily follow - up duration and the average daily follow - up times . Standardize each indicator and calculate the comprehensive work efficiency score of each nurse . The calculation formula is as follows: , where, , , , , , is the weight of each indicator calculated by the Analytic Hierarchy Process (AHP).

[0034] Among them, the standardized number of surgeries, the standardized department coverage, and the participation rate of high - level surgeries respectively represent the standardized scores of nurses in different dimensions. The number of surgeries reflects the workload and surgical experience of nurses. To avoid the absolute value of the number of surgeries having too much impact on the ability index, we standardize the number of surgeries. The department coverage reflects the work experience of nurses in different departments and represents the comprehensive ability of nurses. The participation in high - level surgeries (level - three and level - four surgeries) reflects the ability of nurses to handle complex tasks.

[0035] The calculated ability index (C) is converted into the star rating of nurses according to a preset standard. The specific star - rating criteria are as follows: In addition, the evaluation of nurses' work efficiency aims to quantitatively evaluate the work performance of nurses in the hospital through multi - dimensional work - efficiency indicators. Among them, the total follow - up duration is the total duration of surgeries participated by nurses within a certain period of time, the total number of follow - up times is the total number of surgeries participated by nurses, the average working duration per surgery is the average working duration invested by nurses in each surgery, the number of follow - up days is the number of days when nurses actually participate in surgeries, the average daily follow - up duration is the average duration of surgeries participated by nurses every day, and the average daily follow - up times is the average number of surgeries participated by nurses every day.

[0036] Specifically, to ensure fairness among different indicators, we first standardize all indicators. Using the Min-Max standardization method, we transform the value of each indicator into the range of [0, 1]. The formula is as follows: is the standardized indicator data, is the original data, and are the minimum and maximum values of this indicator respectively. In this way, the dimensions of all ergonomic indicators will be unified, avoiding certain indicators with larger dimensions from dominating the evaluation results.

[0037] The Analytic Hierarchy Process (AHP) is used to calculate the relative importance among indicators and give a preliminary weight assignment. In this embodiment, we invite hospital experts to make pairwise comparisons of the six indicators to obtain a judgment matrix: To ensure the consistency of expert judgments, a consistency ratio (CR) test is performed on the judgment matrix. All CR values are less than 0.1, meeting the requirements of the AHP method. According to the eigenvalue method of AHP, the eigenvector of this matrix is calculated and normalized to obtain the initial weights of each ergonomic indicator. Through calculation, the following initial weights are obtained: Total follow-up duration := 0.43 Total follow-up times := 0.19 Average working duration per operation := 0.12 Follow-up days := 0.06 Average daily follow-up duration := 0.11 Average daily follow-up times := 0.09.

[0038] Furthermore, in step S2, the department resource assessment includes the analysis of the distribution of surgical procedures and surgical grades, as well as the cross-analysis of the capabilities of the primary surgeon and surgical procedures. Among them, the analysis of the distribution of surgical procedures and surgical grades includes: Collect historical data of the operating room, including the operation date, operation grade, name of the surgical procedure, department name, primary surgeon, and participating nurses. Group the operation data of each department according to the operation grade, dynamically detect the surgical grade field (levels 1-4), and count the number of operations of each department at different operation grades; Use a radar chart to display the performance of each department in surgeries of different grades and the strengths of each department in surgeries of different grades. Then stack the number of surgeries of each department according to the operation grade to display the load of each department at different operation grades.

[0039] The cross-analysis of the capabilities of the primary surgeon and surgical procedures further includes: Extract information including the names of doctors participating in the surgery, the types of surgical procedures they are involved in, and the departments where the doctors perform surgeries from the input surgical-related data; By creating a cross-data structure between doctors and surgical procedures, record the types of surgeries each doctor participates in, and match the surgical procedures each doctor is good at with the requirements of the operating room.

[0040] Specifically, in this embodiment, an example data structure: doctor_list = { "Doctor A": {"Surgical Procedure 1","Surgical Procedure 2"}, "Doctor B": {"Surgical Procedure 2","Surgical Procedure 3"}, "Doctor C": {"Surgical Procedure 1","Surgical Procedure 4"} };

[0041] Finally, we integrate the above evaluation results and perform a visual display, which further includes: The ergonomic score of the nurse is a specific value, which is visualized through a radar chart and a violin plot to show the distribution of each ergonomic index and the comprehensive performance of each nurse; The analysis of the distribution of surgical procedures and surgical grades generates a nested data structure, which is adapted to a radar chart and a stacked bar chart. The radar chart shows the performance of each department in surgeries of different grades and the strengths of each department in surgeries of different grades, and the stacked bar chart intuitively shows the load of each department at different surgical grades; The cross-analysis result of the surgeon and surgical procedure capabilities is shown through a cross Venn diagram to show the work distribution of doctors in specific surgical procedures and departments, helping the hospital identify surgical expertise and resource allocation bottlenecks, and optimizing doctor scheduling and resource allocation; In the evaluation of the operating room resource utilization efficiency, the working hours, opening days, and load rate information of the operating room are shown through charts to help managers monitor resource usage in real time. The average working hours of each operating room are shown through a bar chart of the average working hours of each operating room, the opening days of each operating room are shown through a bar chart of the opening days of each operating room within a certain period, the number of doctor working times is shown through a bar chart of the number of doctor working times in each operating room, providing a basis for doctor scheduling and resource optimization. The predicted operating room load curve is shown through a prediction curve and compared with the actual load to help predict future demand.

[0042] Specifically, the ergonomic score of a nurse is a specific value, which represents the work efficiency and performance of the nurse. Tools such as radar charts and violin plots can also be used for visualization to show the distribution of each ergonomic indicator and the comprehensive performance of each nurse. Through the charts, managers can intuitively see the work performance of nurses in different dimensions and the contribution of each indicator to the final score.

[0043] In addition, the cross Venn diagram is a graphical tool for visualizing intersections, unions, and differences between sets. Here, we will use the Venn diagram to show the relationships between different surgical procedures participated by doctors. Each circle in the Venn diagram represents a set (representing a doctor), and the elements inside each circle represent the surgical procedures participated by the doctor. The intersection between circles represents the part where two doctors participate in the same surgical procedure. The cross Venn diagram is generated through the matplotlib_venn library to graphically display each doctor and the surgical procedures they participated in.

[0044] Among them, the set defined as the set of surgical procedures participated by each doctor can be represented as a circle. For example, the set of surgical procedures participated by Doctor A is {Surgical Procedure 1, Surgical Procedure 2}, and the set of surgical procedures participated by Doctor B is {Surgical Procedure 2, Surgical Procedure 3}. Intersection part: The intersection part represents the surgical procedures jointly participated by two doctors. For example, the intersection of Doctor A and Doctor B is Surgical Procedure 2. Difference part: The difference part represents the surgical procedures participated by a certain doctor but not by the other doctor. For example, Doctor A participates in Surgical Procedure 1, but Doctor B does not.

[0045] Through the cross Venn diagram, the hospital can identify the following problems: Surgical expertise bottleneck: The cross Venn diagram helps the hospital identify which doctors have high expertise in certain surgical procedures, while the demand for these surgical procedures in the department is low. For example, Doctor A specializes in Surgical Procedure 1 and Surgical Procedure 2, but the demand for Surgical Procedure 2 is low. Such resource allocation may lead to underutilization of Doctor A's skill resources. By analyzing the cross Venn diagram, hospital managers can identify which doctors' workloads may be affected by low-demand surgical procedures and then optimize them.

[0046] Resource Optimization: Based on the results of the Venn diagram, the hospital can adjust the doctor's scheduling arrangement. Doctors with expertise in certain surgical procedures can be assigned to surgeries with high demand, thereby improving the utilization efficiency of resources. For example, if a doctor only specializes in Surgical Procedure 1 but the demand for Surgical Procedure 1 in the department is not high, then this doctor can be reassigned to a surgical procedure with higher demand. For doctors who only focus on a single surgical procedure, through further training or guidance, their mastery of other surgical procedures can be increased to improve the flexibility of hospital resources and the work efficiency of doctors. The cross Venn diagram not only helps analyze the relationship between doctors and surgical procedures but also helps managers optimize the allocation of department resources. By comparing the surgical expertise of doctors with the department's needs, the hospital can anticipate potential resource bottlenecks in advance and make adjustments. Reasonably arrange doctor scheduling to avoid resource idleness or overcrowding.

[0047] In this embodiment, we display key indicators such as operating room resources, nurse work efficiency, and department surgical grades through the visualization engine module. This module is based on technologies such as WebGL, D3.js, and UpSetJS, providing a real-time interactive and dynamically updated visualization interface for the hospital.

[0048] This module will implement various chart types, such as radar charts, stacked bar charts, cross Venn diagrams, etc., to help hospital managers intuitively analyze the usage of operating room resources, the work efficiency distribution of nurses, and the load situation of the department.

[0049] This visualization module can dynamically update the charts through interactive filtering and control for further analysis of specific data. Among them, the chart rendering is accelerated through WebGL acceleration to handle complex 3D graphics and large-scale data, the data is dynamically refreshed through interactive functions, supporting multiple chart types (such as radar charts, violin charts, stacked bar charts, cross Venn diagrams, etc.), integrating data from multiple dimensions such as operating room resources, nurse work efficiency, and department surgical distribution into a single platform and presenting it through charts.

[0050] In the front-end interface of this embodiment, we use <canvas>Elements are used to create a WebGL rendering context and initialize the WebGL environment. WebGL allows for hardware-accelerated graphics rendering in the browser and is suitable for processing large-scale data sets. This module interacts with the backend through an API interface. The front end fetches data such as hospital operating room resources, nurse work efficiency, and department loads via AJAX requests. The data is transmitted in JSON format for easy visualization processing.

[0051] Specifically, the above embodiments are used in various charts. A radar chart is used to display nurse ability assessments, including the number of surgeries, department coverage, and the participation rate in high-level surgeries, showing the number of surgeries performed by nurses in different departments and the overall ability assessment star rating. WebGL and the D3.js library are used to render the radar chart in real time. It is also used to display nurse work efficiency evaluations, with the input data being the performance of nurses in different areas (such as total follow-up duration, total follow-up times, working hours per surgery, etc.). A radar chart is created using Chart.js, and the data is fetched from the backend via an AJAX request. The numerical values for each dimension are generated through an input form and compared with other nurses, and are displayed on the radar chart. The background color and border of the chart can be customized according to the different levels or roles of the nurses.

[0052] A violin plot will show the distribution of instrument nurses, circulating nurses, and all nurses across multiple work efficiency dimensions. The work efficiency data for each dimension will be shown as an independent "string". Among them, dimension examples: total follow-up duration, total follow-up times, follow-up days, average daily follow-up duration, average daily follow-up times, average working hours per platform surgery. Each part of the violin plot shows the data distribution for that dimension, including: Density distribution: The shape shows the density of the data. The wider the shape, the more nurses there are in that range. Statistical values: Show the minimum, maximum, median, and average values of the data.

[0053] On the front end, visualization libraries such as D3.js, Chart.js, and Plotly.js are used to draw violin plots. Based on the data of each nurse, the chart will show the distribution of all nurses under different ergonomic dimensions. The data distribution of each nurse in each dimension will be displayed in the form of a violin plot. The chart can support interactions such as showing specific values when hovering the mouse and clicking on a certain nurse type to view detailed data. Users can select different time ranges and ergonomic dimensions to dynamically update the data of the violin plot. For example, by selecting a date range (such as from January 1, 2023 to December 31, 2023) or a specific dimension (such as the total follow-up time), the chart will be refreshed in real time to display the relevant data. The width of the violin plot represents the density of the data. The higher the density area, the wider the graph will be shown. For example, if the total follow-up time of most nurses is between 100 hours and 150 hours, the strings in this range will be wider.

[0054] The stacked bar chart shows the surgical distribution of each department under different surgical levels (such as level I, level II, level III, and level IV surgeries). Using the stacked bar chart function of the Chart.js library, the data of different surgical levels are stacked together to clearly show the surgical level distribution of each department. Users can select different time ranges for filtering, and the chart will automatically update to display the relevant data. Set appropriate colors to distinguish different surgical levels (e.g., level I surgery is green, level II surgery is blue, level III surgery is orange, and level IV surgery is red). Each bar represents a department. Each part in the bar represents the number of surgeries at different levels, stacked in order from bottom to top.

[0055] The ordinary bar chart shows the workload of the operating rooms on different dates or time periods, facilitating the analysis of the utilization efficiency of the operating rooms. Use Chart.js to create a standard bar chart, where each bar represents the average working hours or the number of open days of a certain operating room. Users can select a date range, and the chart will be dynamically updated to show the load of each operating room within the selected date interval. The chart can be configured as a vertical or horizontal bar chart. The bars can be distinguished by different colors for different operating rooms. Specific values or labels are also shown when hovering the mouse to increase the readability of the data.

[0056] The application of the cross Venn diagram includes the following steps: Step 1: Select a suitable graph library. To draw a cross Venn diagram, use UpSet.js or a similar graph library to render the data. UpSet.js can effectively display the intersections of sets and provide interactive functions to facilitate users to view the specific content of the data.

[0057] Step 2: Initialize UpSet.js, introduce UpSet.js and necessary CSS styles in the front-end page to ensure the availability of chart rendering. Pass the data passed from the front-end (the intersection of doctors and surgical types) to UpSet.js for rendering. Set the parameters of UpSet.js, such as chart size, color, label display, identification of the intersection area, etc.

[0058] Step 3: Render the cross Venn diagram. Using the above configuration and data, UpSet.js will convert the data into a cross Venn diagram. The intersection part in the diagram will show the overlap of different doctors in surgical types. For example, Doctor A and Doctor C may both be involved in surgeries B and C, but Doctor B is only involved in surgeries A and D.

[0059] Step 4: Add interactive functions to the cross Venn diagram, including: Mouse hover: When the user hovers the mouse over an intersection area, display relevant detailed information, such as doctor name, surgical type and frequency, etc.

[0060] Click on the intersection area: Click on an intersection area to pop up a detailed information window containing the data of that intersection. Show the specific participation of which doctors under this intersection. For example, which surgeries Doctor A participated in and how they were allocated to each surgical level, etc.

[0061] Step 5: Update the chart. During the display of the cross Venn diagram, the user may need to select different time periods and departments to filter the data. By combining with the interactive functions of the front-end, the conditions selected by the user will trigger a new API request. The backend returns the updated data according to the filtering conditions, and the front-end renders the updated cross Venn diagram.

[0062] Second Embodiment This embodiment provides an operation management system for operating room data analysis and visualization, as Figure 2 shown, including: A data preprocessing module, which is used to collect operating room-related data and perform data preprocessing, obtain the actual load of the current operating room according to the operating room-related data, convert the surgical procedure name into a semantic vector through an Embedding model, and realize surgical procedure matching with a standardized procedure library by combining cosine similarity; A multi-dimensional resource evaluation module, which is used to construct an operating room load prediction model based on the actual load of the operating room, trend factors and seasonal adjustment factors, dynamically predict the operating room load in the next time period by adjusting the parameters of the operating room load prediction model, evaluate the operating room resources according to the operating room load prediction result, and then perform nurse resource evaluation and department resource evaluation according to the operating room-related data; An interactive visualization module for integrating the results of the operating room resource assessment, nurse resource assessment, and department resource assessment and visually displaying them, providing an intuitive analysis and real-time interaction for resource load and configuration assessment.

[0063] Among them, the functional implementation of each module in the figure corresponds to each step in the embodiment of the operation management method for operating room data analysis and visualization in the first embodiment, and its functions and implementation processes will not be elaborated here one by one.

[0064] Finally, it should be noted that the above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

[0065] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered that the scope described in this specification is covered.< / canvas>

Claims

1. An operation management method for operating room data analysis and visualization, characterized in that It includes the following steps: S1: Collect relevant operating room data and perform data preprocessing. Obtain the actual load of the current operating room based on the relevant operating room data, convert the surgical procedure name therein into a semantic vector through an Embedding model, and realize surgical procedure matching with a standardized procedure library in combination with cosine similarity; S2: Construct an operating room load prediction model based on the actual load of the operating room, trend factor, and seasonal adjustment factor. Dynamically predict the operating room load in the next time period by adjusting the parameters of the operating room load prediction model, evaluate the operating room resources according to the operating room load prediction result, and then evaluate the nurse resources and department resources according to the relevant operating room data; S3: Integrate the evaluation results of the operating room resources, nurse resources, and department resources and perform visual display to provide an intuitive analysis and real-time interaction for resource load and configuration evaluation.

2. The operation management method for operating room data analysis and visualization according to claim 1, characterized in that In step S1, the collection of relevant operating room data and data preprocessing further includes: Convert the time data in the relevant operating room data into the datetime format for calculating time differences, obtain the duration of each operation, the number of days the operating room is open, and the total working hours of the operating room, and calculate the actual load of the operating room therefrom; Group the surgical procedures in the relevant operating room data according to the procedure code, calculate the average value and standard deviation of the operation duration, table duration, and difference of each procedure, obtain the upper and lower limits of the operation duration, table duration, and difference duration of each procedure according to the 6 Sigma principle, and use the limits to screen abnormal records that do not meet the standards; Establish a mechanism to prevent duplicate name conflicts, and use a combination of name and employee number to ensure the unique identification of each nurse and doctor in the relevant operating room data.

3. The operation management method for operating room data analysis and visualization according to claim 1, characterized in that In step S1, the surgical procedure matching with the standardized procedure library further includes: When entering the surgical procedure name, split the name of the composite surgical procedure through a regular expression; Load the procedure data in the ICD9-CM3 standard library, including the name of each procedure, and use the procedure data as input for pre-training by the Embedding model; Use the pre-trained Embedding model to convert the name of each procedure in the ICD9-CM3 standard library into an embedding vector to generate a set of standard procedure embedding vectors. At the same time, use this pre-trained model to also convert the query procedure input by the user into an embedding vector to obtain a query embedding vector; Calculate the cosine similarity between the query embedding vector and all the embedding vectors in the set of standard procedure embedding vectors, and evaluate the similarity between the query procedure and each standard procedure; Return the standard procedure most similar to the query procedure according to the calculated similarity score.

4. The operation management method for operating room data analysis and visualization according to claim 1, wherein In step S2, constructing an operating room load prediction model based on the actual load of the operating room, trend factor, and seasonal adjustment factor further includes: S21: Obtain the operating room load data for the historical period and ensure consistent time granularity. Decompose the data into a trend term, a seasonal term, and a residual term using the STL decomposition method, and calculate the trend factor using the linear regression method , and obtain the seasonal adjustment factor by quantifying the periodic influence through the seasonal index ; S22: Construct an operating room load prediction model, and the model formula expression is as follows: ， Among them, represents the operating room load in the next period, represents the operating room load in the current period, , , are all empirical parameters that can be adjusted according to historical data, and set the parameters , , initial weights; S23: Train the model, measure the deviation between the predicted value and the actual value using the mean squared error (MSE), and adjust the parameters using the gradient descent method , , , and minimize the MSE to optimize the operating room load prediction model.

5. The operation management method for operating room data analysis and visualization according to claim 4, wherein In step S2, further including dynamically predicting the operating room load in the next time period by adjusting the parameters of the operating room load prediction model: Updating the result influence factor through a sliding window. After adding the actual operating room load data of each time period, recalculate the trend factor and seasonal factor; Adopting incremental training for online learning of the model, and fine-tuning the parameters of the model through exponential weighted adjustment based on new data; Substitute the latest parameters and factors into the operating room load prediction model to predict the operating room load in the next time period.

6. The operation management method for operating room data analysis and visualization according to claim 1, characterized in that, In step S2, the nurse resource assessment includes nurse ability assessment and nurse work efficiency assessment, and further includes: The evaluation of the nurse's ability includes the evaluation of the nurse's ability to participate in the number of surgeries, department coverage, and the participation rate in high-level surgeries. By calculating the nurse ability index to assess the nurse's ability level, and the calculated ability index is converted into the star rating of the nurse according to a preset standard. The calculation formula is as follows: ， Among them, , , are the weight coefficients of the corresponding evaluation dimensions, used to dynamically set and adjust the contributions of different dimensions to the ability index, is the standardized number of surgeries, is the standardized department coverage, is the participation rate in high-level surgeries; Determine the evaluation indicators for the nurses' work efficiency, including the total follow-up duration , the total number of follow-ups , the average working duration per operation , the number of follow-up days , the average follow-up duration per day and the average number of follow-ups per day . Standardize each indicator and calculate the comprehensive work efficiency score for each nurse . The calculation formula is as follows: ， Among them, , , , , , are the weights of each index calculated by the Analytic Hierarchy Process (AHP).

7. The operation management method for operating room data analysis and visualization according to claim 6, characterized in that, In step S2, the department resource assessment includes the analysis of the distribution of surgical procedures and surgical levels, and the cross-analysis of the abilities of the chief surgeons and surgical procedures. Among them, the analysis of the distribution of surgical procedures and surgical levels specifically includes: Collect the historical data of the operating room, including the surgical date, surgical level, name of the surgical procedure, name of the department, chief surgeon, and participating nurses. Group the surgical data of each department according to the surgical level, dynamically detect the surgical level field, and count the number of surgeries of each department at different surgical levels; Use a radar chart to display the performance of each department in surgeries at different levels and the strengths of each department in surgeries at different levels, and then stack the number of surgeries of each department according to the surgical level to display the load of each department at different surgical levels; The cross-analysis of the abilities of the chief surgeons and surgical procedures specifically includes: Extract information including the name of the doctor participating in the surgery, the type of surgical procedure participated by the doctor, and the department of the surgery participated by the doctor from the input surgical-related data; By creating a cross-data structure between doctors and surgical procedures, record the types of surgeries participated by each doctor, and match the surgical procedures that each doctor is good at with the requirements of the operating room.

8. The operation management method for operating room data analysis and visualization according to claim 7, characterized in that, In step S3, further including integrating the evaluation results of the operating room resources, nurse resources, and department resources and performing visual display: The nurse work efficiency score is a specific value, which is visualized through a radar chart and a violin plot to display the distribution of each work efficiency index and the comprehensive performance of each nurse; The analysis of the distribution of surgical procedures and surgical levels generates a nested data structure, which adapts to a radar chart and a stacked bar chart. Use a radar chart to display the performance of each department in surgeries at different levels and the strengths of each department in surgeries at different levels, and use a stacked bar chart to intuitively display the load of each department at different surgical levels; The results of the cross-analysis of the abilities of the chief surgeons and surgical procedures are displayed through a cross Venn diagram to show the work distribution of doctors in specific surgical procedures and departments, helping the hospital identify surgical expertise and resource allocation bottlenecks, and optimize doctor scheduling and resource allocation; In the evaluation of the operating room resource utilization efficiency, the working hours, open days, and load rate information of the operating room are presented through charts to help managers monitor the resource utilization in real time. The average working hours of each operating room are shown through a bar chart of the average working hours of each operating room. The open days of each operating room within a certain period are shown through a bar chart of the open days of each operating room. The number of times each doctor works in each operating room is shown through a bar chart of the number of doctor work times, providing a basis for doctor scheduling and resource optimization. The predicted operating room load curve is shown through a prediction curve and compared with the actual load to help predict future demands.

9. An operation management system for operating room data analysis and visualization, characterized in that, Including: A data preprocessing module for collecting relevant operating room data and performing data preprocessing, obtaining the actual load of the current operating room based on the relevant operating room data, converting the surgical procedure name therein into a semantic vector through an Embedding model, and realizing surgical procedure matching with a standardized surgical procedure library in combination with cosine similarity; A multi-dimensional resource evaluation module for constructing an operating room load prediction model based on the actual load of the operating room, a trend factor, and a seasonal adjustment factor, dynamically predicting the operating room load in the next time period by adjusting the parameters of the operating room load prediction model, evaluating the operating room resources according to the operating room load prediction result, and then evaluating the nurse resources and department resources according to the relevant operating room data; An interactive visualization module for integrating and visually presenting the evaluation results of the operating room resources, nurse resources, and department resources, and providing an intuitive analysis and real-time interaction for the evaluation of the resource load and configuration situation.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the operation management method for operating room data analysis and visualization as described in any one of claims 1-8.

11. An electronic device, characterized in that, Including: One or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the operation management method for operating room data analysis and visualization as described in any one of claims 1-8.

Images