Operation time real-time prediction model based on preoperative and intraoperative information and artificial neural network

Abstract

The invention relates to the technical field of operation time prediction, in particular to an operation time real-time prediction model based on preoperative and intraoperative information and an artificial neural network, which comprises the following steps: a data acquisition module, construction of an artificial neural network model, a data preprocessing module and real-time deployment. The data acquisition module comprises the following steps of collecting a large amount of performed operation data in advance, including preoperative information, intraoperative information and actual operation time, classifying operation types and difficulty levels, classifying physical states of operation doctors, and performing preoperative acquisition; data of age, gender, weight, health condition, basic disease and previous operation records of a patient are subjected to collection model selection, a neural network is fed forward, and a simple multi-layer sensor is adopted; according to the scheme, the preoperative physical condition of the patient is compared with the previous operation record, the experience of an operation doctor is collected, and the sudden unexpected condition in the operation is predicted, so that the operation time is predicted and evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field of operation time prediction, and in particular to a real-time operation time prediction model based on preoperative and intraoperative information and an artificial neural network. Background Art

[0002] In today's context of increasingly tight medical resources, the rapid development of medical technology and the massive accumulation of medical data, traditional surgical prediction methods have gradually revealed their limitations. They are often difficult to fully predict surgical times and usually rely on doctors' experience or rule-based algorithms. These rules may be summarized from historical data, but they are often not flexible enough and have difficulty adapting to complex clinical changes. Because these models lack the ability to learn automatically, their prediction accuracy is limited when dealing with complex cases. In addition, data differences between different hospitals and different types of surgeries may also lead to poor generalization of the model, making it difficult to make effective predictions in different scenarios.

[0003] A search revealed Chinese patent application CN118735015A, a method and electronic device for predicting surgical duration. The invention uses an electronic device to first obtain a video file of the current surgical procedure and, based on the video file, determine the target surgical stage to which the current surgical procedure belongs. Different surgical stages have different predicted durations, so the electronic device can determine the predicted duration of the target surgical stage based on the predicted durations of each surgical stage. The electronic device then updates the predicted duration of the target surgical stage and the total duration of the surgical type to which the target surgical stage belongs based on the actual duration of the target surgical stage.

[0004] A method and electronic device for predicting operation duration makes predictions based on previous operation durations and operation stages. Traditional analysis systems cannot predict operation duration based on the patient's physical condition, doctor's experience, and doctor's physical state, which affects the accuracy of operation time prediction. To this end, the present invention proposes a real-time operation time prediction model based on preoperative and intraoperative information and artificial neural networks. By collecting the patient's physical condition, doctor's experience, and doctor's physical state from multiple perspectives and using an artificial neural network model to perform preoperative prediction and evaluation on the collected information, the accuracy of operation duration prediction is improved, and combined with real-time data during surgery, it helps doctors to reasonably arrange the operation process. Summary of the Invention

[0005] Technical problems solved: It is impossible to collect the patient's preoperative physical condition, compare the patient's preoperative physical condition with historical surgical records, and reasonably predict the operation duration based on the surgical experience and physical condition of different surgeons.

[0006] In response to the deficiencies of the prior art, the present invention provides a real-time prediction model for surgical time based on preoperative and intraoperative information and artificial neural networks, thereby solving the technical problems mentioned in the background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] The method includes the following steps: data acquisition module, building artificial neural network model, data preprocessing module, and real-time deployment;

[0009] The data acquisition module includes the following steps: S1: collecting a large amount of surgical data in advance, including preoperative information, intraoperative information and actual surgical time. Based on past surgical records, it can help doctors understand the patient's specific conditions in advance, analyze a large amount of surgical data in advance, and help predict the actual time of the surgery;

[0010] Collect a large amount of surgeons' preoperative and intraoperative physical conditions in advance, and make different surgery duration predictions based on different surgeons' physical conditions, thereby improving surgery quality, ensuring patient safety, and optimizing work arrangements;

[0011] Classify the type and difficulty of surgery, classify the physical condition of the surgeon, and predict the duration of surgery based on the difficulty of the surgery to reduce the prediction error;

[0012] Preoperative data collection: The patient's age, gender, weight, health status, underlying diseases, and previous surgical records are collected. Based on the previous surgical records, the patient is predicted to have complications during surgery and a preliminary estimate of the duration of the surgery is made based on the complications.

[0013] Intraoperative data collection: whether the surgery went according to plan, and whether any complications or complications were encountered;

[0014] The incision time, blood loss, suturing time, equipment usage, anesthesia response, and the proportion of time in different stages of the operation were collected.

[0015] Electronic devices are worn by surgeons and patients to collect data on the surgeon's physical condition during surgery, monitor the patient's vital signs and surgical progress, and collect data on sudden changes in the patient's bleeding volume, anesthesia response, etc.

[0016] The construction of the artificial neural network model includes the following steps,S2: model selection, feedforward neural network, simple multi-layer perceptron (MLP), comparing the predicted value and the true value, calculating the loss, and reducing the error;

[0017] Convolutional neural network, the input data contains spatial structure information; CNN can be used to extract features;

[0018] Long-short-term memory networks process time series data and capture the temporal information of real-time intraoperative data. Ensemble learning combines multiple models to improve prediction accuracy, explore potential connections between data, and provide comprehensive information support for subsequent predictions and decision-making.

[0019] Model training learns the complex relationship between preoperative and intraoperative information and surgical time. Cross-validation helps evaluate the model on different data, reduces the bias caused by uneven data distribution, and accurately predicts surgical duration.

[0020] Model evaluation and prediction accuracy assessment: Based on common evaluation indicators including mean square error and mean absolute error, the model should be evaluated within a reasonable time frame;

[0021] Real-time assessment, based on historical surgeries combined with the patient's physical condition, and dynamic adjustment to predict the final operation time based on actual progress.

[0022] In a possible implementation, the data preprocessing module includes the following steps:

[0023] S3: Fill in missing values ​​in the data to keep the data as complete as possible during model training, standardize the data to improve the model's convergence process and training efficiency, and encode categorical variables to ensure that each category is processed independently in the model.

[0024] In a possible implementation, the real-time deployment includes the following steps: S4: deploying the trained model into a real-time surgical environment of a hospital;

[0025] Real-time data input: real-time data during surgery and surgical steps are used as input. The model can update the surgery time prediction in real time based on this information and dynamically adjust the prediction results.

[0026] Surgical time predictions can be used as a reference for doctors and family members, helping doctors rationally plan surgical procedures, equipment scheduling, and postoperative care. Family members can also use this time prediction to reduce long periods of idle waiting and reduce anxiety about the unknown.

[0027] After the operation is completed, the neural network model is further optimized by comparing the predicted time with the actual operation time.

[0028] Beneficial effects compared with existing technologies:

[0029] 1. In this solution, the data collection module is used to collect and store past surgical records, and classify the surgical types and difficulty levels. The patient's preoperative physical condition, the presence of underlying diseases, and past surgeries are recorded, so that the surgery required by the patient can be compared with the surgery recorded in the past, so as to make a preliminary prediction of the patient's surgical duration based on the past surgical records. At the same time, the patient's sudden unexpected conditions during the operation can be predicted based on the past surgical records. The surgeon's surgical experience and physical condition on the day are recorded, and the duration of operations of different types and difficulties are predicted based on surgeons with different experiences. In this way, the accuracy of the surgical duration prediction is improved. By recording the patient's intraoperative information, the surgical time prediction is updated according to the patient's real-time intraoperative data information, and the prediction results are dynamically adjusted.

[0030] 2. This solution uses an artificial neural network model to compare predicted and true values ​​of the detected data and calculate losses. The input data contains spatial structural information, processes time series data, and captures the timing information of real-time intraoperative data to improve prediction accuracy. The model performs prediction evaluation based on common evaluation indicators. It also dynamically adjusts the actual progress based on historical patient surgeries and the patient's physical condition to predict the final surgery time. By filling in missing values ​​and standardizing the data, the accuracy of surgery duration prediction is increased.

[0031] 3. In this solution, by deploying the trained model into the hospital's real-time surgical environment, combined with real-time intraoperative data and the surgical process, the remaining surgical time can be predicted in real time, and family members can be notified promptly. Accurate time expectations can reduce family members' anxiety about the unknown and reduce long periods of idle waiting. At the same time, based on the predicted end time of the operation, relevant follow-up steps can be prepared in advance for patients and families. The surgical time prediction can be used as a reference for doctors to help them reasonably arrange the surgical process, make equipment scheduling and postoperative care. After the operation is completed, the predicted time is compared with the actual surgical time, and the neural network model is further optimized to improve the accuracy of the model's subsequent prediction of surgical time. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention with reference to the accompanying drawings.

[0033] Figure 1 It is a schematic diagram of the process structure of the present invention; DETAILED DESCRIPTION

[0034] In the description of the present invention, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "axial", "radial", "circumferential", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0035] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0036] In the present invention, unless otherwise clearly specified and limited, terms such as "install", "connect", "connect", and "fix" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integrated connection; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances.

[0037] In the present invention, unless otherwise clearly specified and limited, the first feature "above" or "below" the second feature may be that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. In the description of this specification, the descriptions with reference to the terms "one scheme", "some schemes", "examples", "specific examples", or "some examples" etc. mean that the specific features, structures, materials or characteristics described in conjunction with the scheme or example are included in at least one scheme or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same scheme or example. Moreover, the specific features, structures, materials or characteristics described may be combined in an appropriate manner in any one or more schemes or examples.

[0038] In order to more clearly and completely illustrate the technical solution of the present invention, the present invention will be further described below with reference to the accompanying drawings.

[0039] Example 1:

[0040] Please refer to Figure 1As shown, this embodiment introduces a real-time prediction model for operation time based on preoperative and intraoperative information and artificial neural network, which includes the following steps: data acquisition module, construction of artificial neural network model, data preprocessing module, and real-time deployment;

[0041] The data acquisition module includes the following steps: collecting a large amount of surgical data that has been performed in advance, including preoperative information, intraoperative information and actual surgical time, so that doctors can understand the patient's specific situation in advance based on past surgical records. The surgical plan customized according to past surgical records can help reduce sudden problems during the operation and improve the success rate of the operation. At the same time, by analyzing a large amount of surgical data that has been performed, it can help predict the actual time of the operation, ensure patient safety, and provide a strong reference basis for future operations; collecting a large amount of physical conditions of surgeons before and during the operation in advance, making different predictions based on the physical conditions of different doctors, helping hospitals to optimize surgical processes, improve surgical quality, and ensure patient safety from multiple aspects. Through data analysis, hospitals can understand the physical conditions of doctors in real time, thereby optimizing work arrangements and avoiding surgical risks caused by doctors' health problems; classifying the types and difficulty levels of operations, and classifying the physical conditions of surgeons, and can predict the duration of the operation based on the difficulty of the operation, thereby reducing the error value of the predicted time;

[0042] Preoperative data collection: The patient's age, gender, weight, health status, underlying diseases, and previous surgical records are collected. Based on the collected preoperative examination results of the patient, including blood tests, imaging tests, and electrocardiograms, it helps predict possible complications during the operation. Combined with the surgeon's surgical experience and physical condition, the operation duration is predicted, thereby improving the accuracy of the operation duration prediction;

[0043] During the operation, data is collected to determine whether the operation is carried out according to the plan and whether there are any complex situations or complications. The data are collected during the preparation stage, the incision stage, the suturing stage, etc. At the same time, electronic devices are worn on the surgeon and the patient. The electronic devices collect the surgeon's physical condition during the operation and detect the surgeon's physical fitness and concentration during the operation. According to the surgeon's own physical fitness and the medical team's initial prediction, the electronic devices worn by the patient can monitor the patient's vital signs and the progress of the operation to ensure the safety and success of the operation. The current operation process is recorded by the electronic device, and the patient's incision time and bleeding amount, suturing time, and equipment usage are recorded. , anesthesia reaction and the time proportion of different stages are collected, and the target surgical stage to which the current surgical process belongs is determined based on the video file. Whether it will affect the predicted time, and the predicted duration of different surgical stages is different, and then the electronic equipment can determine the predicted duration of the target surgical stage according to the predicted duration of each surgical stage, and update the predicted total duration in real time according to the actual intraoperative information, thereby improving the utilization rate of the operating room, and archiving the collected information to help subsequent similar surgeries to make preliminary surgical time predictions and data analysis, and to find the key factors affecting the surgical time, so as to facilitate the subsequent improvement of the efficiency of surgical arrangements through the established prediction model;

[0044] The construction of the artificial neural network model includes the following steps, S2: model selection, feedforward neural network, simple multi-layer perceptron (MLP), compare the predicted value and the true value, calculate the loss, and reduce the error; convolutional neural network, input data contains spatial structure information; CNN can be used to extract features; long short-term memory network, process time series data, capture the time series information of real-time data during surgery; ensemble learning, combining multiple models to improve prediction accuracy. The large model can integrate and analyze these different types of data, explore the potential connections between the data, and provide comprehensive information support for subsequent prediction and decision-making;

[0045] Model training learns the complex relationship between preoperative and intraoperative information and surgical time. Cross-validation can help evaluate the model's performance on different datasets and reduce bias caused by uneven data partitioning, enabling accurate prediction of surgical duration.

[0046] Model evaluation and prediction accuracy assessment: Preliminary predictions are made based on common evaluation metrics such as mean square error and mean absolute error to achieve a reasonable timeframe. The final predicted surgery time is then dynamically adjusted based on historical surgeries, the patient's physical condition, and actual progress.

[0047] The data preprocessing module includes the following steps: S3 fills in missing values ​​in the data to maintain data integrity, avoid data loss or sample shortages due to missing values, prevent machine learning models from generating errors due to missing data, and ensure that the data during model training is as complete as possible; standardize the data to keep all features on the same scale, prevent certain features from affecting model training due to excessively large or small dimensions, and improve the model's convergence process and training efficiency;

[0048] Encoding categorical variables enables the model to understand categorical variables and use their information for prediction, while avoiding the false order relationship problem that may be caused by label encoding and ensuring that each category is treated independently in the model;

[0049] Real-time deployment includes the following steps: S4: Deploy the trained model to the hospital's real-time surgical environment, and predict the remaining operation time in real time during the operation. Real-time data input, including the real-time data of the surgical steps during the operation, is used as input. Combined with the real-time data during the operation, if there are any unexpected situations during the operation, the model can update the operation time prediction in real time based on this information, and the prediction results can be dynamically adjusted; the operation time prediction can be used as a reference for doctors and patients' families, helping doctors to reasonably arrange the operation process, make equipment scheduling and postoperative care. Family members can reduce their anxiety about the unknown through time expectations, and at the same time reduce long periods of idle waiting;

[0050] After the operation is completed, the neural network model is further optimized by comparing the predicted time with the actual operation time to improve the accuracy of the model's subsequent prediction of operation time.

[0051] Finally, it should be noted that the above embodiments are merely examples for the purpose of illustrating the present invention and are not intended to limit the embodiments. Those skilled in the art will readily appreciate that other variations or modifications based on the above description are possible. It is not necessary and impossible to provide an exhaustive list of all embodiments. However, obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A real-time prediction model for surgical time based on preoperative and intraoperative information and artificial neural networks, characterized by: The following steps are involved: Data acquisition module, building artificial neural network model, data preprocessing module, real-time deployment; The data acquisition module includes the following steps: S1: collecting a large amount of surgical data in advance, including preoperative information, intraoperative information and actual surgical time. Based on past surgical records, it can help doctors understand the patient's specific conditions in advance, analyze a large amount of surgical data in advance, and help predict the actual time of the surgery; Collect a large amount of surgeons' preoperative and intraoperative physical conditions in advance, and make different surgery duration predictions based on different surgeons' physical conditions, thereby improving surgery quality, ensuring patient safety, and optimizing work arrangements; Classify the type and difficulty of surgery, classify the physical condition of the surgeon, and predict the duration of surgery based on the difficulty of the surgery to reduce the prediction error; Preoperative data collection: The patient's age, gender, weight, health status, underlying diseases, and previous surgical records are collected. Based on the previous surgical records, the patient is predicted to have complications during surgery and a preliminary estimate of the duration of the surgery is made based on the complications. Intraoperative data collection: whether the surgery went according to plan, and whether any complications or complications were encountered; The incision time, blood loss, suturing time, equipment usage, anesthesia response, and the proportion of time in different stages of the operation were collected. Electronic devices are worn by surgeons and patients to collect data on the surgeon's physical condition during surgery, monitor the patient's vital signs and surgical progress, and collect data on sudden changes in the patient's bleeding volume, anesthesia response, etc. The construction of the artificial neural network model includes the following steps,S2: model selection, feedforward neural network, simple multi-layer perceptron (MLP), comparing the predicted value and the true value, calculating the loss, and reducing the error; Convolutional neural network, the input data contains spatial structure information; CNN can be used to extract features; Long-short-term memory networks process time series data and capture the temporal information of real-time intraoperative data. Ensemble learning combines multiple models to improve prediction accuracy, explore potential connections between data, and provide comprehensive information support for subsequent predictions and decision-making. Model training learns the complex relationship between preoperative and intraoperative information and surgical time. Cross-validation helps evaluate the model on different data, reduces the bias caused by uneven data distribution, and accurately predicts surgical duration. Model evaluation and prediction accuracy assessment: Based on common evaluation indicators including mean square error and mean absolute error, the model should be evaluated within a reasonable time frame; Real-time assessment, based on historical surgeries combined with the patient's physical condition, and dynamic adjustment to predict the final operation time based on actual progress.

2. The real-time prediction model for surgical time based on preoperative and intraoperative information and artificial neural network according to claim 1, characterized in that: The data preprocessing module includes the following steps: S3: Fill in missing values ​​in the data to keep the data as complete as possible during model training, standardize the data to improve the model's convergence process and training efficiency, and encode categorical variables to ensure that each category is processed independently in the model.

3. The real-time prediction model for surgical time based on preoperative and intraoperative information and artificial neural network according to claim 1, characterized in that: The real-time deployment includes the following steps: S4: deploying the trained model into a real-time surgical environment in a hospital; Real-time data input: real-time data during surgery and surgical steps are used as input. The model can update the surgery time prediction in real time based on this information and dynamically adjust the prediction results. Surgical time predictions can be used as a reference for doctors and family members, helping doctors rationally plan surgical procedures, equipment scheduling, and postoperative care. Family members can also use this time prediction to reduce long periods of idle waiting and reduce anxiety about the unknown. After the operation is completed, the neural network model is further optimized by comparing the predicted time with the actual operation time.

Images