Personalized anesthesia management method and system

Abstract

The invention provides a personalized anesthesia management method and system. The method comprises the following steps: collecting real-time physiological data flow and previous medical record information of a patient to preliminarily evaluate operation requirements, and generating a preliminary anesthesia scheme; performing simulation prediction processing on the individualized reaction mode through a state-action value function updating mechanism by applying a deep Q network algorithm, and optimizing anesthetic selection and dosage by adopting a dynamic adaptive dosage adjustment technology to generate an optimized anesthetic scheme; the method comprises the following steps: extracting continuous physiological state parameters of a patient by applying an adaptive feedback control algorithm, presetting a safety threshold value, performing real-time comparison and analysis through a real-time parameter adjustment mechanism, performing automatic early warning processing on a trend deviating from a normal range by adopting an abnormal trend early warning technology, and generating a real-time adjustment instruction; and based on the real-time adjustment instruction, obtaining a patient recovery trajectory and subjective comfort evaluation, performing retrospective analysis, and generating a personalized anesthesia management scheme. According to the technical scheme provided by the invention, the safety of personalized anesthesia is remarkably improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of anesthesia management technology, and in particular, to a personalized anesthesia management method and system. Background Art

[0002] With the continuous development of medical technology and the increasing demand of patients for personalized medical services, in modern surgical operations, anesthesia management not only needs to consider the patient's current physiological state, but also needs to be comprehensively evaluated in combination with their previous medical records. Traditional anesthesia plans are often based on standardized drug doses and processes, which makes it difficult to fully consider the individual differences of each patient. Therefore, there is an urgent need for a method that can collect patient physiological data streams in real time and combine them with previous medical records to generate a personalized preliminary anesthesia plan. In addition, the selection and dosage of anesthetic drugs need to be dynamically adjusted during the anesthesia process to adapt to the patient's changing physiological state and ensure stability and safety during surgery. After surgery, a retrospective analysis of the anesthesia effect is also required to optimize future anesthesia decisions.

[0003] At present, most hospitals use an experience-based anesthesia program formulation method, that is, the anesthesiologist selects anesthetic drugs and their dosages based on the patient's age, weight, medical history and other basic information, and refers to the established anesthesia guidelines. Although this method ensures the safety of anesthesia to a certain extent, it lacks the ability to dynamically respond to the patient's real-time physiological changes. In recent years, some medical institutions have begun to introduce computer-assisted systems and use machine learning algorithms to make preliminary predictions on anesthetic drug combinations, but most of these systems remain at the static analysis level and fail to achieve true real-time adjustment and personalized management.

[0004] The main drawback of existing anesthesia management programs is that traditional experience-based anesthesia programs usually require determining the anesthesia plan based on respiratory and monitoring indicators based on empirical values, and doctors need to pay continuous attention to the indication information, resulting in low anesthesia management efficiency. Summary of the invention

[0005] The embodiments of the present application provide a personalized anesthesia management method and system to solve the problem of low anesthesia management efficiency in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a personalized anesthesia management method, comprising:

[0007] Collecting the patient's real-time physiological data stream and previous medical record information, conducting a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generating a preliminary anesthesia plan;

[0008] Using a deep Q-network algorithm, through the state-action value function update mechanism in the deep Q-network algorithm, simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan, and use dynamic adaptive dose adjustment technology to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation and prediction results to generate an optimized anesthesia plan;

[0009] Using an adaptive feedback control algorithm, extracting the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, presetting a safety threshold for the patient's continuous physiological state parameters, performing a real-time comparative analysis of the patient's continuous physiological state parameters and the safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm, using an abnormal trend early warning technology, automatically performing early warning processing on the trend of the real-time comparative analysis results that deviates from the normal range, and generating a real-time adjustment instruction;

[0010] Based on the real-time adjustment instructions, the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed are obtained, and the decision points in the process of optimizing the anesthesia plan are retrospectively analyzed and processed to generate a personalized anesthesia management plan.

[0011] Optionally, the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through the state-action value function update mechanism in the deep Q network algorithm, and the dynamic adaptive dose adjustment technology is used to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation prediction processing results to generate an optimized anesthesia plan, including:

[0012] Extracting anesthetic drugs and dosage information in the preliminary anesthesia plan, classifying and normalizing the anesthetic drugs and dosage information, and generating a standardized anesthetic drug configuration table;

[0013] Based on the standardized anesthetic drug configuration table, a deep Q-network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through a state-action value function update mechanism in the deep Q-network algorithm, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination;

[0014] Based on the optimized anesthetic drug combination, a dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing result, so as to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan and generate a drug dose adjustment result;

[0015] Based on the drug dosage adjustment result, all optimization parameters in the state-action value function optimization process are integrated to generate an optimized anesthesia plan.

[0016] Optionally, based on the standardized anesthetic drug configuration table, a deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through a state-action value function update mechanism in the deep Q network algorithm, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination, including:

[0017] By means of a feature extraction method, the pharmacological mechanism features in the standardized anesthetic drug configuration table are extracted, the patient personalized features in the previous medical record information are extracted, the interaction effect between the pharmacological mechanism features and the patient personalized features is calculated, and an anesthetic drug combination feature matrix is ​​generated;

[0018] Based on the anesthetic drug combination feature matrix, a deep Q-network algorithm is used to simulate and predict the individualized reaction patterns of different anesthetic drug combinations in the preliminary anesthesia plan through a state-action value function update mechanism in the deep Q-network algorithm to generate an individualized reaction prediction pattern;

[0019] Based on the individualized response prediction model, by introducing a reward mechanism in iterative learning, a positive reward is given to the drug combination that achieves the expected anesthetic effect and has the least side effects among the different anesthetic drug combinations, so as to gradually optimize the state-action value function and generate an optimized state-action value function;

[0020] The optimal anesthetic drug selection and dosage information in the optimized state-action value function is extracted, and the effect verification processing is performed on the optimal anesthetic drug selection and dosage information to generate an optimized anesthetic drug combination.

[0021] Optionally, based on the optimized anesthetic drug combination, a dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing result, so as to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan and generate a drug dose adjustment result, including:

[0022] Normalizing the optimized anesthetic drug combination, quantifying the interactions between the drugs in the optimized anesthetic drug combination, and generating an anesthetic drug combination effect matrix;

[0023] Using dynamic adaptive dose adjustment technology, the anesthetic drug combination effect matrix is ​​input into the pharmacokinetic model in the dynamic adaptive dose adjustment technology, and the anesthetic effect and side effect risk in the simulation prediction processing result are quantitatively evaluated to generate a quantitative evaluation result;

[0024] Obtaining an optimization decision basis based on the quantitative evaluation result, optimizing the anesthetic drug selection and dosage in the preliminary anesthesia plan according to the optimization decision basis, and generating an optimized anesthetic drug selection and dosage;

[0025] Based on the optimized anesthetic drug selection and dosage, the drug combination effect in the anesthetic drug selection and dosage optimization process in the preliminary anesthesia plan is evaluated to generate a drug dosage adjustment result.

[0026] Optionally, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparative analysis is performed between the patient's continuous physiological state parameters and the safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm. An abnormal trend early warning technology is used to automatically warn the trend of the real-time comparative analysis result that deviates from the normal range, and a real-time adjustment instruction is generated, including:

[0027] Extracting the time-series physiological data during the implementation of the optimized anesthesia scheme, performing time series analysis on the time-series physiological data, recording the patient's physiological dynamic changes during the time series analysis, and generating a dynamic file of the patient's physiological state;

[0028] Based on the patient's physiological state dynamic file, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparison and analysis is performed between the patient's continuous physiological state parameters and the safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a preset safety threshold range;

[0029] Based on the preset safety threshold range, an abnormal trend warning technology is used to set up a warning mechanism, and according to the warning mechanism, automatic warning processing is performed on the trend that deviates from the normal range in the real-time comparative analysis to generate an automatic warning result;

[0030] The abnormal trend in the automatic warning result is recorded, and the anesthesia depth in the optimized anesthesia scheme is adjusted in real time according to the abnormal trend, and a real-time adjustment instruction is generated.

[0031] Optionally, based on the patient's physiological state dynamic profile, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time parameter adjustment mechanism in the adaptive feedback control algorithm is used to perform a real-time comparative analysis between the patient's continuous physiological state parameters and the safety threshold to generate a preset safety threshold range, including:

[0032] Performing correlation analysis based on the dynamic change parameters in the patient's physiological state dynamic profile, identifying potential synergistic effects in the correlation analysis results, and generating a potential synergistic effect report;

[0033] Based on the potential synergistic effect report, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparative analysis is performed between the patient's continuous physiological state parameters and the preset safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a real-time comparative analysis result;

[0034] Based on the real-time comparison data in the real-time comparison analysis result and the individual differences of the patients, dynamically adjusting the preset safety threshold to adapt to the fluctuation of the real-time comparison data and the individual differences of the patients, and generating a dynamic safety threshold;

[0035] The patient's continuous physiological state parameter is compared with the dynamic safety threshold item by item. When the patient's continuous physiological state parameter is greater than the dynamic safety threshold, the upper limit of the dynamic safety threshold is increased to expand the dynamic safety threshold range to generate a preset safety threshold range.

[0036] Optionally, based on the real-time adjustment instruction, obtaining the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, retrospectively analyzing and processing the decision points in the process of optimizing the anesthesia plan, and generating a personalized anesthesia management plan, including:

[0037] Extract all adjustment decision points based on the real-time adjustment instruction, classify all the adjustment decision points, perform logical relationship analysis on different decision point categories in the classification results, and generate a decision point category file;

[0038] Matching different decision point categories in the decision point category file to the optimized anesthesia plan to clarify the implementation process of the optimized anesthesia plan, obtaining the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, and generating a recovery trajectory and comfort evaluation report;

[0039] Performing retrospective analysis on the turning points of the recovery trajectory and the dynamic changes of the comfort evaluation in the recovery trajectory and comfort evaluation report to generate a retrospective analysis report;

[0040] The improvement suggestions in the retrospective analysis report are summarized, and the improvement suggestions are introduced into the real-time adjustment instructions to optimize the real-time adjustment instructions in a targeted manner and generate a personalized anesthesia management plan.

[0041] In a second aspect, an embodiment of the present application provides a personalized anesthesia management system, including:

[0042] A collection module, used to collect the patient's real-time physiological data stream and previous medical record information, conduct a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generate a preliminary anesthesia plan;

[0043] A processing module is used to perform simulation and prediction processing on the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan based on the preliminary anesthesia plan by using a deep Q-network algorithm and a state-action value function update mechanism in the deep Q-network algorithm, and to optimize the selection and dosage of anesthetic drugs in the preliminary anesthesia plan according to the simulation and prediction processing results by using a dynamic adaptive dose adjustment technology to generate an optimized anesthesia plan;

[0044] An analysis module, for extracting the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme by using an adaptive feedback control algorithm based on the optimized anesthesia scheme, presetting safety thresholds for the patient's continuous physiological state parameters, performing real-time comparative analysis between the patient's continuous physiological state parameters and the preset safety thresholds through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm, automatically performing early warning processing on the trend that deviates from the normal range in the real-time comparative analysis by using an abnormal trend early warning technology, and generating real-time adjustment instructions;

[0045] A generation module is used to obtain the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed based on the real-time adjustment instructions, retrospectively analyze and process the decision points in the process of optimizing the anesthesia plan, and generate a personalized anesthesia management plan.

[0046] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a personalized anesthesia management method as described in the first aspect.

[0047] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a personalized anesthesia management method as described in the first aspect.

[0048] In an embodiment of the present application, the real-time physiological data stream and the previous medical record information of the patient are collected, and a preliminary assessment of the surgical needs is performed based on the real-time physiological data stream and the previous medical record information to generate a preliminary anesthesia plan; a deep Q network algorithm is used, and the state-action value function update mechanism in the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan, and a dynamic adaptive dose adjustment technology is used to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation prediction processing results to generate an optimized anesthesia plan; an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, and a safety threshold is preset for the patient's continuous physiological state parameters. Through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm, the patient's continuous physiological state parameters and the safety threshold are compared and analyzed in real time, and the abnormal trend warning technology is used to automatically warn the trend that deviates from the normal range in the real-time comparison and analysis results, and a real-time adjustment instruction is generated; based on the real-time adjustment instruction, the patient's recovery trajectory and subjective comfort evaluation after the completion of the optimized anesthesia plan are obtained, and the decision points in the process of optimizing the anesthesia plan are retrospectively analyzed and processed to generate a personalized anesthesia management plan. By collecting the patient's real-time physiological data stream and previous medical history information, combined with the preliminary assessment of surgical needs, a preliminary anesthesia plan is generated to ensure the individualization and pertinence of the anesthesia plan. The deep Q network algorithm is used to simulate and predict the individualized response pattern of different anesthetic drug combinations, and the dynamic adaptive dose adjustment technology is used to optimize the selection and dose of anesthetic drugs, thereby improving the safety and effectiveness of anesthesia. Furthermore, the patient's continuous physiological state parameters during the implementation process are monitored and compared in real time through an adaptive feedback control algorithm, and abnormal trend warning technology is used to achieve automatic warning processing, ensuring the safety of the anesthesia process. Finally, based on the real-time adjustment instructions, the patient's recovery trajectory and subjective comfort evaluation are retrospectively analyzed to generate a personalized anesthesia management plan, realizing refined management of the entire process from pre-anesthesia to post-anesthesia.

[0049] Furthermore, by extracting the anesthetic drugs and dosage information in the preliminary anesthesia plan, classifying and normalizing it, a standardized anesthetic drug configuration table is generated to ensure the basic scientificity and consistency of anesthetic drug selection. Based on this configuration table, the state-action value function update mechanism of the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination. On this basis, the dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk, and finally all optimization parameters are integrated to generate an optimized anesthesia plan, which greatly improves the accuracy and personalization of the anesthesia plan and reduces the potential anesthesia risk.

[0050] Furthermore, by analyzing the time series of the sequential physiological data during the implementation of the optimized anesthesia plan, the patient's physiological dynamic changes are recorded, and a dynamic file of the patient's physiological state is generated, which provides a solid data foundation for subsequent real-time monitoring. Based on this file, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters, preset safety thresholds and conduct real-time comparative analysis to generate a preset safety threshold range. By setting up an early warning mechanism, the trend that deviates from the normal range is automatically warned, and an automatic warning result is generated. The depth of anesthesia is adjusted in real time according to the abnormal trend, and a real-time adjustment instruction is generated. This method not only enhances the real-time monitoring capability during anesthesia, but also improves the response speed and accuracy to emergencies, ensuring the safety and stability of the anesthesia process.

[0051] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A flowchart of a personalized anesthesia management method provided in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of the structure of a personalized anesthesia management system provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0057] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0059] Figure 1 A flowchart of a personalized anesthesia management method is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0060] 101. Collect the patient's real-time physiological data stream and previous medical record information, conduct a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generate a preliminary anesthesia plan;

[0061] In this step, the real-time physiological data stream refers to the patient's physiological parameters continuously collected during the anesthesia process, such as heart rate, blood pressure, blood oxygen saturation, etc. These data are transmitted in real time through various monitoring devices to reflect the patient's current physiological state.

[0062] Past medical records include the patient's age, weight, medical history, allergies, and previous anesthesia experience, which provides important background information about the patient's health status and potential risks.

[0063] The preliminary assessment of surgical needs is a process of analyzing the patient's specific surgical needs based on real-time physiological data streams and previous medical history information. The assessment aims to determine the type of anesthesia and drug combination that is most suitable for the patient and generate a preliminary anesthesia plan.

[0064] The preliminary anesthesia plan refers to the initial anesthesia plan developed based on the above assessment results, including the selection of specific anesthetic drugs and their doses to meet surgical needs and ensure patient safety.

[0065] In the embodiments of the present application, first, real-time physiological data from different monitoring devices are integrated through multi-source data fusion technology; second, a comprehensive evaluation is performed using a machine learning algorithm in combination with the patient's previous medical history information; third, based on the evaluation results, a personalized preliminary anesthesia plan is generated; finally, this plan is provided to the anesthesiologist as a reference to ensure the individualization and targeting of the anesthesia plan.

[0066] Suppose a male patient needs to undergo heart bypass surgery. His medical records show that he has a history of hypertension and diabetes, so he needs to be prepared for anesthesia before the operation. First, the medical team uses a variety of monitoring devices to collect the patient's real-time physiological data such as heart rate, blood pressure, and blood oxygen saturation; second, combined with the patient's medical record information, the machine learning algorithm analyzes the patient's specific situation; third, based on the analysis results, a preliminary anesthesia plan containing appropriate anesthetic drugs and their dosages is generated; finally, the anesthesiologist prepares the anesthetic drugs required for the operation according to this plan to ensure the safety and effectiveness of the anesthesia process.

[0067] 102. Using a deep Q network algorithm, through the state-action value function update mechanism in the deep Q network algorithm, simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan, and use dynamic adaptive dose adjustment technology to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation prediction results to generate an optimized anesthesia plan;

[0068] In this step, the deep Q network algorithm is a reinforcement learning method that optimizes the selection and dosage of anesthetic drugs by simulating and predicting the effects of different anesthetic drug combinations. It can dynamically adjust decisions to adapt to different patient response patterns.

[0069] The state-action value function update mechanism refers to a mechanism in the deep Q-network algorithm that is used to continuously update and optimize the value estimate of taking a certain action in each state. This mechanism helps the model better understand which anesthetic drug combinations are most beneficial to patients.

[0070] Individualized response patterns refer to the fact that different patients may have different physiological responses to the same combination of anesthetic drugs. This pattern reflects the individual differences of patients and is crucial to achieving personalized anesthesia management.

[0071] Dynamic adaptive dose adjustment technology is a method of adjusting the dose of anesthetic drugs based on simulated predicted treatment results to ensure optimal anesthetic effect and minimal side effects.

[0072] The optimized anesthesia plan refers to the final anesthesia plan after optimization by the deep Q network algorithm. This plan takes into account the patient's individualized response pattern to ensure the safety and effectiveness of the anesthesia process.

[0073] In the embodiments of the present application, first, the anesthetic drugs and dosage information in the preliminary anesthesia plan are extracted, classified and normalized; secondly, based on the standardized configuration table, the state-action value function update mechanism of the deep Q network algorithm is utilized to simulate and predict the effects of different anesthetic drug combinations; thirdly, the anesthetic effect and side effect risk are quantitatively evaluated through dynamic adaptive dose adjustment technology; finally, all optimization parameters are integrated to generate an optimized anesthesia plan.

[0074] For example, continuing with the previous example, assume that a preliminary anesthesia plan has been generated through preliminary evaluation. First, the medical team extracts the anesthetic drugs and dosage information in the plan, and classifies and normalizes it; second, based on the standardized configuration table, the deep Q network algorithm is used to simulate and predict the effects of different anesthetic drug combinations; third, the dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthesia effect and side effect risk; finally, all optimization parameters are integrated to generate an optimized anesthesia plan to ensure the safety and effectiveness of the anesthesia process.

[0075] 103. Using an adaptive feedback control algorithm, extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, preset a safety threshold for the patient's continuous physiological state parameters, perform real-time comparative analysis between the patient's continuous physiological state parameters and the safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm, use abnormal trend early warning technology to automatically warn of trends that deviate from the normal range in the real-time comparative analysis results, and generate real-time adjustment instructions;

[0076] In this step, the adaptive feedback control algorithm is an algorithm used to monitor and adjust the patient's continuous physiological state parameters in real time. It can automatically adjust the dosage of anesthetic drugs according to real-time data to ensure that the patient is always within a safe range.

[0077] The patient's continuous physiological state parameters refer to the patient's physiological indicators that are continuously monitored during anesthesia, such as electrocardiogram, electroencephalogram, etc. These parameters reflect the patient's physiological changes during anesthesia.

[0078] The preset safety threshold refers to the safety range set for each physiological parameter. When the actual parameter deviates from this range, the system will trigger the early warning mechanism.

[0079] Real-time comparative analysis refers to comparing the physiological parameters collected in real time with the preset safety thresholds to identify any trends that deviate from the normal range.

[0080] Abnormal trend early warning technology is a technology used to detect and warn of abnormal physiological trends, which can detect potential problems at an early stage and take measures to avoid adverse consequences.

[0081] Real-time adjustment instructions refer to adjustment suggestions generated based on abnormal trend warning results, which guide anesthesiologists to adjust the dosage or type of anesthetic drugs in a timely manner to ensure patient safety.

[0082] In the embodiments of the present application, first, the temporal physiological data in the process of implementing the optimized anesthesia scheme is extracted and time series analysis is performed; secondly, based on the analysis results, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters, and a safety threshold is preset; thirdly, a real-time parameter adjustment mechanism is used to perform a real-time comparative analysis of the patient's continuous physiological state parameters and the safety threshold; finally, the abnormal trend warning technology is used to automatically warn of trends that deviate from the normal range and generate real-time adjustment instructions.

[0083] For example, continuing with the above example, assume that the optimized anesthesia plan has been implemented. First, the medical team extracts the time-series physiological data during the implementation process and performs time series analysis; second, based on the analysis results, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters and preset safety thresholds; third, through the real-time parameter adjustment mechanism, the patient's continuous physiological state parameters are compared and analyzed with the safety threshold in real time; finally, the abnormal trend warning technology is used to automatically warn the trend that deviates from the normal range, and generate real-time adjustment instructions to ensure the safety and stability of the anesthesia process.

[0084] 104. Based on the real-time adjustment instruction, the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed are obtained, and the decision points in the process of optimizing the anesthesia plan are retrospectively analyzed and processed to generate a personalized anesthesia management plan.

[0085] In this step, the recovery trajectory refers to the process of the patient recovering from the anesthesia state after the end of anesthesia. This trajectory reflects the effect of the anesthetic drug and the patient's recovery speed.

[0086] Subjective comfort evaluation refers to the patient's perception of their own comfort level after anesthesia, which is usually obtained through questionnaires or other methods.

[0087] Retrospective decision point analysis is a review of key decision points during the anesthesia process with the goal of evaluating the effectiveness of each decision and drawing lessons from it.

[0088] Personalized anesthesia management plan refers to the final anesthesia management strategy generated based on the results of retrospective analysis. This plan is not only applicable to the current patient, but also provides a reference for similar cases in the future.

[0089] In the embodiments of the present application, first, the patient's recovery trajectory and subjective comfort evaluation are obtained after the optimized anesthesia scheme is completed; second, a retrospective analysis is performed on the decision points during the anesthesia process; third, the effectiveness of each decision is evaluated, and lessons learned are summarized; finally, a personalized anesthesia management plan is generated to provide guidance for future anesthesia management.

[0090] For example, continuing with the previous example, assume that the optimized anesthesia plan has been completed. First, the medical team obtains the patient's postoperative recovery trajectory and subjective comfort evaluation; second, a retrospective analysis of the decision points during the anesthesia process is performed; third, the effectiveness of each decision is evaluated and lessons learned are summarized; finally, a personalized anesthesia management plan is generated to ensure the safety and effectiveness of the anesthesia process and provide guidance for future anesthesia management.

[0091] In order to solve the complexity and personalized needs in the process of anesthetic drug selection and dosage optimization, in some embodiments, the use of a deep Q-network algorithm to optimize the anesthesia scheme in step 102 includes: extracting anesthetic drugs and dosage information in the preliminary anesthesia scheme, classifying and normalizing the anesthetic drugs and dosage information, and generating a standardized anesthetic drug configuration table; based on the standardized anesthetic drug configuration table, using a deep Q-network algorithm, through the state-action value function update mechanism in the deep Q-network algorithm, simulating and predicting the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia scheme, and optimizing the state-action value function through iterative learning to generate an optimized anesthetic drug combination; based on the optimized anesthetic drug combination, using a dynamic adaptive dose adjustment technology to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing results, so as to optimize the anesthetic drug selection and dose in the preliminary anesthesia scheme and generate a drug dose adjustment result; based on the drug dose adjustment result, integrating all optimization parameters in the state-action value function optimization process to generate an optimized anesthesia scheme.

[0092] In this embodiment, the standardized anesthetic drug configuration table refers to a data table generated by classifying and normalizing the anesthetic drugs and dosage information in the preliminary anesthesia plan.

[0093] The state-action value function update mechanism is a core mechanism in the deep Q-network algorithm, which is used to continuously update and optimize the value estimate of taking a certain action in each state. This mechanism helps the model better understand which anesthetic drug combinations are most beneficial to patients and gradually improves the prediction accuracy through iterative learning.

[0094] Individualized response patterns refer to the different physiological responses that different patients may have to the same combination of anesthetic drugs. This pattern reflects the individual differences of patients and is crucial to achieving personalized anesthesia management. Through simulation and predictive processing, these response patterns can be captured more accurately, thereby optimizing the anesthesia plan.

[0095] Dynamic adaptive dose adjustment technology is a method of adjusting the dose of anesthetic drugs based on simulated predicted treatment results to ensure optimal anesthetic effect and minimal side effects. This technology is based on quantitative evaluation of anesthetic effect and side effect risk, and can adjust drug doses in real time to adapt to the patient's specific conditions.

[0096] The drug dose adjustment result refers to the anesthetic drug dose recommendation after optimization by dynamic adaptive dose adjustment technology. It is generated based on the simulation prediction processing results and aims to ensure the safety and effectiveness of the anesthesia process.

[0097] The optimized anesthesia plan refers to the final anesthesia plan generated after integrating all optimized parameters. This plan takes into account the patient's individualized response pattern to ensure the safety and effectiveness of the anesthesia process.

[0098] In the embodiments of the present application, first, the anesthetic drugs and dosage information in the preliminary anesthesia plan are extracted, classified and normalized, and a standardized anesthetic drug configuration table is generated; secondly, based on this configuration table, the state-action value function update mechanism of the deep Q network algorithm is utilized to simulate and predict the effects of different anesthetic drug combinations, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination; thirdly, the dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing results, optimize the anesthetic drug selection and dosage, and generate the drug dosage adjustment results; finally, all optimization parameters in the state-action value function optimization process are integrated to generate the final optimized anesthesia plan.

[0099] Here is a specific example:

[0100] For example, suppose an anesthesiologist needs to perform abdominal surgery anesthesia on a 45-year-old female patient whose medical record shows a history of asthma. First, the medical team extracts the anesthetic drugs and dosage information in the preliminary anesthesia plan based on the patient's previous medical records and real-time physiological data streams, and classifies and normalizes them to generate a standardized anesthetic drug configuration table; secondly, based on this configuration table, the state-action value function update mechanism of the deep Q network algorithm is used to simulate and predict the effects of different anesthetic drug combinations, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination; thirdly, the dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing results, optimize the anesthetic drug selection and dosage, and generate drug dose adjustment results; finally, all optimization parameters in the state-action value function optimization process are integrated to generate the final optimized anesthesia plan to ensure the safety and effectiveness of the anesthesia process, and special consideration is given to the patient's asthma history to avoid the use of drugs that may cause asthma symptoms.

[0101] In order to solve the complexity and personalized needs in the process of anesthetic drug selection and dosage optimization, in some embodiments, the optimization of the anesthesia scheme based on the standardized anesthetic drug configuration table in step 102 includes: extracting the pharmacological mechanism characteristics in the standardized anesthetic drug configuration table through a feature extraction method, extracting the patient's personalized characteristics in the previous medical record information, calculating the interaction effect between the pharmacological mechanism characteristics and the patient's personalized characteristics, and generating an anesthetic drug combination feature matrix; based on the anesthetic drug combination feature matrix, using a deep Q network algorithm, through the state-action value function update mechanism in the deep Q network algorithm, the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia scheme are simulated and predicted to generate an individualized response prediction model; based on the individualized response prediction model, by introducing a reward mechanism in iterative learning, a positive reward is given to the drug combination that achieves the expected anesthetic effect and has the least side effects among the different anesthetic drug combinations, so as to gradually optimize the state-action value function and generate an optimized state-action value function; extracting the best anesthetic drug selection and dosage information in the optimized state-action value function, performing effect verification processing on the best anesthetic drug selection and dosage information, and generating an optimized anesthetic drug combination.

[0102] In this embodiment, the pharmacological mechanism characteristics refer to the unique mechanism of action of each anesthetic drug, such as the target of action, metabolic pathway, etc. These characteristics determine how the drug affects the patient's physiological state and are an important basis for selecting appropriate anesthetic drugs.

[0103] Patient personalized characteristics refer to data about individual differences of patients extracted from previous medical records, including age, weight, medical history, allergy records, etc. These characteristics reflect the different reaction patterns that patients may have to anesthetic drugs and are crucial to achieving personalized anesthesia management.

[0104] The interaction effect refers to the interaction between the pharmacological mechanism characteristics and the patient's personalized characteristics. By calculating this interaction effect, the individualized response pattern of different anesthetic drug combinations for specific patients can be more accurately predicted, thereby optimizing the anesthesia regimen.

[0105] The anesthetic drug combination feature matrix is ​​a data structure that integrates pharmacological mechanism characteristics and patient personalized characteristics. It is used to represent the potential effects of different anesthetic drug combinations. It provides input data for the deep Q-network algorithm to help the model better understand which drug combinations are most beneficial to patients.

[0106] Individualized response prediction model refers to the simulation and prediction of the response pattern of different anesthetic drug combinations to specific patients through deep Q-network algorithm. This model reflects the physiological changes that may occur in patients when receiving different anesthetic drug combinations, which helps to optimize the selection and dosage of anesthetic drugs.

[0107] The reward mechanism is a method used in reinforcement learning that guides the learning direction of the model by giving positive or negative feedback.

[0108] The optimized state-action value function refers to the best state-action value estimate obtained after iterative learning. It can more accurately reflect the value of taking a certain action under different states and help the model make better decisions.

[0109] The optimal anesthetic drug selection and dosage information refers to the optimal anesthetic drug combination and its dosage recommendation extracted based on the optimized state-action value function. It is the core content of the final optimized anesthesia plan, which aims to ensure the safety and effectiveness of the anesthesia process.

[0110] Effectiveness verification processing is the testing and evaluation of the optimal anesthetic drug selection and dosage information before actual application to ensure its safety and effectiveness.

[0111] Optimizing the anesthetic drug combination refers to the best anesthetic drug selection and dosage recommendation extracted based on the optimized state-action value function. It is the core content of the final generated anesthesia plan and aims to ensure the safety and effectiveness of the anesthesia process.

[0112] In the embodiments of the present application, first, a feature extraction method is used to extract pharmacological mechanism features from a standardized anesthetic drug configuration table, and individual patient features are extracted from previous medical records, and the interaction effect of the two is calculated to generate an anesthetic drug combination feature matrix; secondly, based on this feature matrix, the state-action value function update mechanism of the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations to specific patients, and generate an individualized response prediction model; thirdly, a reward mechanism is introduced in iterative learning to give positive rewards to drug combinations that achieve the expected anesthetic effect with minimal side effects, and gradually optimize the state-action value function; finally, the optimal anesthetic drug selection and dosage information in the optimized state-action value function is extracted, and the effect is verified to generate an optimized anesthetic drug combination.

[0113] Here is a specific example:

[0114] For example, suppose an anesthesiologist needs to perform anesthesia for a 60-year-old male patient undergoing hip replacement surgery, whose medical record shows a history of hypertension and chronic obstructive pulmonary disease. First, the medical team uses feature extraction methods to extract pharmacological mechanism features from the standardized anesthetic drug configuration table, extracts patient personalized features from previous medical record information, and calculates the interaction effect between the two to generate an anesthetic drug combination feature matrix; secondly, based on this feature matrix, the state-action value function update mechanism of the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations to specific patients, and generate an individualized response prediction model; thirdly, a reward mechanism is introduced in iterative learning to give positive rewards to drug combinations that achieve the expected anesthetic effect with minimal side effects, and gradually optimize the state-action value function; finally, the optimal anesthetic drug selection and dosage information in the optimized state-action value function is extracted, and the effect is verified to generate an optimized anesthetic drug combination to ensure the safety and effectiveness of the anesthesia process, and special consideration is given to the patient's history of hypertension and chronic obstructive pulmonary disease to avoid the use of drugs that may aggravate the condition.

[0115] In order to address the complexity and personalized needs in the process of anesthetic drug selection and dosage optimization, in some embodiments, the further optimization based on the optimized anesthetic drug combination described in step 102 includes: normalizing the optimized anesthetic drug combination, quantifying the interactions between the drugs in the optimized anesthetic drug combination, and generating an anesthetic drug combination effect matrix; using dynamic adaptive dose adjustment technology, inputting the anesthetic drug combination effect matrix into the pharmacokinetic model in the dynamic adaptive dose adjustment technology, and quantitatively evaluating the anesthetic effect and side effect risk in the simulation prediction processing results to generate a quantitative evaluation result; obtaining an optimization decision basis based on the quantitative evaluation result, and optimizing the anesthetic drug selection and dosage in the preliminary anesthesia plan according to the optimization decision basis to generate an optimized anesthetic drug selection and dosage; based on the optimized anesthetic drug selection and dosage, evaluating the drug combination effect in the anesthetic drug selection and dosage optimization process in the preliminary anesthesia plan to generate a drug dosage adjustment result.

[0116] In this embodiment, the normalization process refers to converting the dose and characteristics of each drug in the optimized anesthetic drug combination to the same scale to ensure that the comparison between different drugs is comparable.

[0117] Interaction quantification refers to the quantitative analysis of the synergistic or antagonistic effects between drugs in an optimized anesthetic drug combination. By calculating these interactions, the overall effect of the drug combination can be predicted more accurately, thereby avoiding potential risks.

[0118] The anesthetic drug combination effect matrix is ​​a data structure that integrates the interactions between various drugs. It is used to represent the expected effects of different anesthetic drug combinations. It provides input data for dynamic adaptive dose adjustment technology, helping the model to better evaluate anesthetic effects and side effect risks.

[0119] The pharmacokinetic model is a mathematical model that describes the absorption, distribution, metabolism and excretion process of drugs in the body. By inputting the anesthetic drug combination effect matrix into the model, the actual effect of the drug combination can be simulated and predicted, providing a basis for quantitative evaluation.

[0120] Quantitative evaluation results refer to the data generated after quantitative evaluation of anesthetic effects and side effect risks through pharmacokinetic models. It reflects the potential impact of different anesthetic drug combinations on patients and provides a basis for optimizing decision-making.

[0121] The basis for optimized decision-making refers to the decision recommendations generated based on quantitative evaluation results, which provide guidance on how to optimize the selection and dosage of anesthetic drugs. It comprehensively considers the anesthetic effect and the risk of side effects, and aims to ensure the safety and effectiveness of the anesthesia process.

[0122] Optimizing the selection and dosage of anesthetic drugs refers to the best anesthetic drugs and their dosage recommendations adjusted according to the optimization decision basis. It is the core content of the final optimized anesthesia plan and aims to ensure the safety and effectiveness of the anesthesia process.

[0123] The evaluation of drug combination effects is the process of verifying the effects of drug combinations during the selection of anesthetic drugs and dose optimization in the preliminary anesthesia plan. In this way, the anesthesia plan can be further optimized, potential risks can be reduced, and reliable data support can be provided for clinical applications.

[0124] The drug dose adjustment result refers to the anesthetic drug dose recommendation after optimization by dynamic adaptive dose adjustment technology. It is generated based on quantitative evaluation results and optimized decision-making basis, and aims to ensure the safety and effectiveness of the anesthesia process.

[0125] In the embodiments of the present application, first, the optimized anesthetic drug combination is normalized, the interactions between the drugs are quantified, and an anesthetic drug combination effect matrix is ​​generated; secondly, a dynamic adaptive dose adjustment technique is used to input the anesthetic drug combination effect matrix into a pharmacokinetic model, and the anesthetic effect and side effect risk in the simulation prediction processing results are quantitatively evaluated to generate a quantitative evaluation result; thirdly, an optimization decision basis is obtained based on the quantitative evaluation result, and the anesthetic drug selection and dosage in the preliminary anesthesia plan are optimized according to the optimization decision basis to generate an optimized anesthetic drug selection and dosage; finally, based on the optimized anesthetic drug selection and dosage, the drug combination effect in the anesthetic drug selection and dosage optimization process in the preliminary anesthesia plan is evaluated to generate a drug dosage adjustment result.

[0126] Here is a specific example:

[0127] For example, suppose an anesthesia operating room needs to perform anesthesia for a male patient who has a brain tumor resection surgery and whose medical record shows a history of epilepsy and renal insufficiency. First, the medical team normalizes the optimized anesthetic drug combination, quantifies the interactions between the drugs, and generates an anesthetic drug combination effect matrix; secondly, the dynamic adaptive dose adjustment technology is used to input the anesthetic drug combination effect matrix into the pharmacokinetic model, and the anesthetic effect and side effect risk in the simulation prediction processing results are quantitatively evaluated to generate quantitative evaluation results; thirdly, based on the quantitative evaluation results, the optimization decision basis is obtained, and the anesthetic drug selection and dosage in the preliminary anesthesia plan are optimized according to the optimization decision basis to generate the optimized anesthetic drug selection and dosage; finally, based on the optimized anesthetic drug selection and dosage, the drug combination effect in the anesthetic drug selection and dosage optimization process in the preliminary anesthesia plan is evaluated to generate drug dosage adjustment results to ensure the safety and effectiveness of the anesthesia process, and special consideration is given to the patient's history of epilepsy and renal insufficiency to avoid the use of drugs that may aggravate the condition.

[0128] In order to solve the problem of real-time monitoring and dynamic adjustment of the patient's physiological state during anesthesia, in some embodiments, the use of the adaptive feedback control algorithm for real-time comparative analysis and early warning processing in step 103 includes: extracting the time-series physiological data during the implementation of the optimized anesthesia scheme, performing time series analysis on the time-series physiological data, recording the dynamic changes of the patient's physiology during the time series analysis, and generating a dynamic file of the patient's physiological state; based on the dynamic file of the patient's physiological state, using the adaptive feedback control algorithm to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme, presetting a safety threshold for the patient's continuous physiological state parameters, and performing real-time comparative analysis between the patient's continuous physiological state parameters and the safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a preset safety threshold range; based on the preset safety threshold range, using abnormal trend early warning technology to set an early warning mechanism, automatically warning the trend that deviates from the normal range in the real-time comparative analysis according to the early warning mechanism, and generating an automatic early warning result; recording the abnormal trend in the automatic early warning result, and adjusting the anesthetic depth in the optimized anesthesia scheme in real time according to the abnormal trend, and generating a real-time adjustment instruction.

[0129] In this embodiment, the time series physiological data refers to the time series of the patient's physiological parameters continuously collected during the anesthesia process, such as heart rate, blood pressure, blood oxygen saturation, etc.

[0130] Time series analysis is a statistical method used to identify and predict patterns and trends in time series data. By performing time series analysis on temporal physiological data, we can capture the changing patterns of the patient's physiological state and generate a dynamic profile of the patient's physiological state.

[0131] The patient's physiological status dynamic profile is a data record generated based on the results of time series analysis, reflecting the patient's physiological changes throughout the anesthesia process.

[0132] The preset safety threshold refers to the safety range set for each physiological parameter. When the actual parameter deviates from this range, the system will trigger the early warning mechanism.

[0133] The real-time parameter adjustment mechanism is part of the adaptive feedback control algorithm, which is used to dynamically adjust the anesthetic drug dosage or other related parameters according to real-time data. It ensures that the patient's physiological state always remains within the preset safety range.

[0134] Abnormal trend early warning technology is a technology used to detect and warn of abnormal physiological trends. It can discover potential problems at an early stage and take measures to avoid adverse consequences. It identifies trends that deviate from the normal range by setting up an early warning mechanism.

[0135] Automatic warning results refer to warning information generated by abnormal trend warning technology, which prompts anesthesiologists to pay attention to changes in specific physiological parameters. These results help the medical team take timely action to prevent potential risks.

[0136] Real-time adjustment instructions refer to adjustment suggestions generated based on automatic warning results, which guide anesthesiologists to adjust the dosage or type of anesthetic drugs in a timely manner to ensure patient safety.

[0137] In the embodiments of the present application, first, the time-series physiological data in the process of implementing the optimized anesthesia scheme are extracted, and time series analysis is performed on these data, the dynamic changes of the patient's physiology in the time series analysis process are recorded, and a dynamic file of the patient's physiological state is generated; secondly, based on this dynamic file, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters, and a safety threshold is preset, and these parameters are compared and analyzed with the safety threshold in real time through a real-time parameter adjustment mechanism to generate a preset safety threshold range; thirdly, based on the preset safety threshold range, an abnormal trend warning technology is used to set an early warning mechanism, and the trend that deviates from the normal range in the real-time comparison analysis is automatically warned according to the early warning mechanism, and an automatic early warning result is generated; finally, the abnormal trend in the automatic early warning result is recorded, and the anesthetic depth in the optimized anesthesia scheme is adjusted in real time according to the abnormal trend, and a real-time adjustment instruction is generated.

[0138] Here is a specific example:

[0139] For example, suppose an anesthesiologist needs to perform anesthesia for a male patient who has a history of liver cirrhosis and chronic kidney disease in his medical record. First, the medical team extracts the temporal physiological data during the implementation of the optimized anesthesia plan, performs time series analysis on these data, records the patient's physiological dynamic changes during the time series analysis, and generates a dynamic file of the patient's physiological state; second, based on this dynamic file, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters, and a safety threshold is preset. These parameters are compared and analyzed with the safety threshold in real time through the real-time parameter adjustment mechanism to generate a preset safety threshold range; third, based on the preset safety threshold range, an abnormal trend warning technology is used to set an early warning mechanism, and the trend that deviates from the normal range in the real-time comparison analysis is automatically warned according to the early warning mechanism to generate an automatic early warning result; finally, the abnormal trend in the automatic early warning result is recorded, and the anesthetic depth in the optimized anesthesia plan is adjusted in real time according to the abnormal trend, and a real-time adjustment instruction is generated to ensure the safety and effectiveness of the anesthesia process, and the patient's history of liver cirrhosis and chronic kidney disease are specially considered to avoid the use of drugs that may aggravate the condition.

[0140] In order to solve the problem of real-time monitoring and dynamic adjustment of the patient's physiological state during anesthesia, in some embodiments, the further optimization based on the patient's physiological state dynamic file in step 103 includes: performing correlation analysis based on the dynamically changing parameters in the patient's physiological state dynamic file, identifying potential synergistic effects in the correlation analysis results, and generating a potential synergistic effect report; based on the potential synergistic effect report, using an adaptive feedback control algorithm to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, presetting a safety threshold for the patient's continuous physiological state parameters, and performing real-time comparative analysis on the patient's continuous physiological state parameters and the preset safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a real-time comparative analysis result; based on the real-time comparative data in the real-time comparative analysis result and the individual differences of the patients, dynamically adjusting the preset safety threshold to adapt to the fluctuation of the real-time comparative data and the individual differences of the patients, and generating a dynamic safety threshold; comparing the patient's continuous physiological state parameters with the dynamic safety threshold item by item, and when the patient's continuous physiological state parameters are greater than the dynamic safety threshold, increasing the upper limit of the dynamic safety threshold to expand the dynamic safety threshold range, and generating a preset safety threshold range.

[0141] In this embodiment, the patient's physiological status dynamic profile is a data record generated based on the time series analysis result, reflecting the physiological changes of the patient during the entire anesthesia process.

[0142] Dynamically changing parameters refer to physiological parameters that fluctuate over time in the patient's physiological status dynamic file, such as heart rate, blood pressure, blood oxygen saturation, etc. The change patterns of these parameters can reflect the patient's immediate physiological condition.

[0143] Correlation analysis is a statistical method used to identify the relationships between different dynamically changing parameters. Through this analysis, potential synergistic effects can be discovered, that is, the interactions between certain physiological parameters may affect the overall anesthetic effect.

[0144] The potential synergistic effect report refers to a report generated based on the results of correlation analysis, which describes the potential synergistic effects between different physiological parameters. This report helps to understand which combinations of physiological parameters may have a significant impact on the anesthetic effect.

[0145] The preset safety threshold refers to the safety range set for each physiological parameter. When the actual parameter deviates from this range, the system will trigger an early warning mechanism. These thresholds are determined based on clinical experience and historical data to ensure the safety of anesthesia for patients.

[0146] The real-time parameter adjustment mechanism is part of the adaptive feedback control algorithm, which is used to dynamically adjust the anesthetic drug dosage or other related parameters according to real-time data. It ensures that the patient's physiological state always remains within the preset safety range.

[0147] Real-time comparative analysis results refer to the results generated by real-time comparison of the patient's continuous physiological state parameters with the preset safety thresholds through the real-time parameter adjustment mechanism. These results help the medical team take timely action to prevent potential risks.

[0148] The dynamic safety threshold refers to the safety threshold that is dynamically adjusted based on the real-time comparison data in the real-time comparison analysis results and the individual differences of patients. It can adapt to the patient's individualized response pattern and ensure the safety and effectiveness of the anesthesia process.

[0149] The dynamic safety threshold range refers to the generation of a more flexible safety range by adjusting the upper and lower limits of the dynamic safety threshold to adapt to fluctuations in real-time comparison data and individual differences among patients.

[0150] In an embodiment of the present application, first, a correlation analysis is performed based on the dynamically changing parameters in the patient's physiological state dynamic file, potential synergistic effects in the correlation analysis results are identified, and a potential synergistic effect report is generated; secondly, based on the potential synergistic effect report, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, preset safety thresholds for these parameters, and perform real-time comparative analysis through a real-time parameter adjustment mechanism to generate real-time comparative analysis results; thirdly, based on the real-time comparative data in the real-time comparative analysis results and the individual differences of patients, the preset safety threshold is dynamically adjusted to adapt to the fluctuations of the real-time comparative data and the individual differences of patients, and a dynamic safety threshold is generated; finally, the patient's continuous physiological state parameters are compared with the dynamic safety threshold item by item. When the patient's continuous physiological state parameters are greater than the dynamic safety threshold, the upper limit of the dynamic safety threshold is increased to expand the dynamic safety threshold range, and a preset safety threshold range is generated.

[0151] Here is a specific example:

[0152] For example, suppose an anesthesiologist needs to perform anesthesia for spinal correction surgery, and his medical record shows a history of severe asthma. First, the medical team conducts correlation analysis based on the dynamic change parameters in the patient's physiological state dynamic file, identifies the potential synergistic effects in the correlation analysis results, and generates a potential synergistic effect report; second, based on the potential synergistic effect report, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, preset safety thresholds for these parameters, and perform real-time comparative analysis through the real-time parameter adjustment mechanism to generate real-time comparative analysis results; third, based on the real-time comparative data and individual differences of patients in the real-time comparative analysis results, the preset safety threshold is dynamically adjusted to adapt to the fluctuation of real-time comparative data and individual differences of patients, and a dynamic safety threshold is generated; finally, the patient's continuous physiological state parameters are compared with the dynamic safety threshold item by item. When the patient's continuous physiological state parameters are greater than the dynamic safety threshold, the upper limit of the dynamic safety threshold is increased to expand the dynamic safety threshold range, and a preset safety threshold range is generated to ensure the safety and effectiveness of the anesthesia process, and the patient's asthma history is specially considered to avoid the use of drugs that may cause asthma symptoms.

[0153] In order to solve the problems of retrospective analysis and personalized management during anesthesia, in some embodiments, the optimization based on the real-time adjustment instruction in step 104 includes: extracting all adjustment decision points based on the real-time adjustment instruction, classifying all the adjustment decision points, performing logical relationship analysis on different decision point categories in the classification results, and generating a decision point category file; matching different decision point categories in the decision point category file to the optimized anesthesia plan to clarify the implementation process of the optimized anesthesia plan, obtaining the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, and generating a recovery trajectory and comfort evaluation report; retrospectively analyzing and processing the turning points of the recovery trajectory and the dynamic changes of the comfort evaluation in the recovery trajectory and comfort evaluation report to generate a retrospective analysis report; summarizing the improvement suggestions in the retrospective analysis report, and introducing the improvement suggestions in the real-time adjustment instruction to optimize the real-time adjustment instruction in a targeted manner to generate a personalized anesthesia management plan.

[0154] In this embodiment, the real-time adjustment instruction refers to an adjustment suggestion generated based on the comparative analysis results of the patient's continuous physiological state parameters and the preset safety thresholds, which guides the anesthesiologist to adjust the dosage or type of anesthetic drugs in a timely manner to ensure patient safety.

[0155] The adjustment decision point refers to the time point during the anesthesia process when a specific adjustment decision is made based on real-time data and algorithm feedback.

[0156] Logical relationship analysis is a method used to identify the interrelationships between different decision point categories. It helps the medical team understand the cause-effect relationship and synergy between various decision points, thereby improving the accuracy of anesthesia management.

[0157] The decision point category file is a data record generated based on the classification results and logical relationship analysis. It describes in detail the characteristics and interrelationships of different types of adjustment decision points, and provides basic data support for subsequent retrospective analysis.

[0158] Subjective comfort evaluation refers to the patient's feeling of comfort after anesthesia, which is usually obtained through questionnaires or other methods. This helps to understand the impact of the anesthesia process on patients and evaluate their satisfaction.

[0159] The recovery trajectory and comfort evaluation report is a comprehensive report generated based on the patient's recovery trajectory and subjective comfort evaluation, reflecting the recovery status and patient feelings after anesthesia.

[0160] Retrospective analysis is the process of in-depth analysis of key data in the recovery trajectory and comfort evaluation report. It summarizes the successful experiences and shortcomings of the anesthesia process by identifying the turning points of the recovery trajectory and the dynamic changes of the comfort evaluation.

[0161] The retrospective analysis report refers to a summary document generated by analyzing the recovery trajectory and comfort evaluation report, which puts forward improvement suggestions and future optimization directions. This report provides a scientific basis for personalized anesthesia management.

[0162] A personalized anesthesia management plan refers to the final anesthesia management strategy generated by combining improvement suggestions and real-time adjustment instructions. This plan is not only applicable to current patients, but also provides a reference for similar cases in the future, ensuring that each patient can get the anesthesia plan that best suits them.

[0163] In an embodiment of the present application, first, all adjustment decision points are extracted based on the real-time adjustment instructions, these decision points are classified, and the logical relationship analysis of different decision point categories in the classification results is performed to generate a decision point category file; secondly, the different decision point categories in the decision point category file are matched to the optimized anesthesia plan to clarify the implementation process of the optimized anesthesia plan, obtain the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, and generate a recovery trajectory and comfort evaluation report; thirdly, the turning points of the recovery trajectory and the dynamic changes of the comfort evaluation in the recovery trajectory and comfort evaluation report are retrospectively analyzed and processed to generate a retrospective analysis report; finally, the improvement suggestions in the retrospective analysis report are summarized, and the real-time adjustment instructions are optimized in a targeted manner by introducing the improvement suggestions in the real-time adjustment instructions to generate a personalized anesthesia management plan.

[0164] Here is a specific example:

[0165] For example, suppose an anesthesiologist needs to perform anesthesia for a heart valve replacement surgery, and his medical record shows a history of arrhythmia. First, the medical team extracts all adjustment decision points based on the real-time adjustment instructions, classifies these decision points, and performs logical relationship analysis on the different decision point categories in the classification results to generate a decision point category file; secondly, the different decision point categories in the decision point category file are matched to the optimized anesthesia plan to clarify the implementation process of the optimized anesthesia plan, obtain the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, and generate a recovery trajectory and comfort evaluation report; thirdly, the turning point of the recovery trajectory and the dynamic changes of the comfort evaluation in the recovery trajectory and comfort evaluation report are retrospectively analyzed and processed to generate a retrospective analysis report; finally, the improvement suggestions in the retrospective analysis report are summarized, and the real-time adjustment instructions are optimized in a targeted manner by introducing improvement suggestions in the real-time adjustment instructions, and a personalized anesthesia management plan is generated to ensure the safety and effectiveness of the anesthesia process, and the patient's history of arrhythmia is specially considered to avoid the use of drugs that may aggravate the condition.

[0166] The present application considers that in order to solve the complexity and personalized demand problems in the anesthetic drug selection and dosage optimization process in the prior art, the invention embodiment proposes this optional solution to solve the technical problem of personalized anesthesia management, and thus proposes a new optional solution, which includes:

[0167] Based on the anesthetic drug combination feature matrix, a deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through the state-action value function update mechanism in the deep Q network algorithm to generate an individualized response prediction pattern, including:

[0168] Based on the anesthetic drug combination feature matrix, the different scale features in the anesthetic drug combination feature matrix are standardized by standardization processing, the main feature vectors in the standardized result are extracted by principal component analysis to reduce the data dimension and retain significant information, and the association rule mining technology is used to identify the association degree of the significant information to generate the basic state-action value;

[0169] The basic state-action value is calculated using the following formula:

[0170]

[0171] Among them, Q b(s,a) is the basic state-action value for taking action a in the current state; s is the current state, which describes the patient's physiological parameters and environmental conditions during anesthesia; a is the current action, which indicates the anesthetic drug combination and its dosage selected in the current state; α is the learning rate, which determines the balance between the weight of new information and the weight of old information; r is the immediate reward, which indicates the direct benefit or effect brought by the current decision; γ is the discount factor, which is used to measure the importance of future rewards; s' is the new state after the action is executed, which reflects the patient's latest physiological parameters and environmental conditions after receiving a specific anesthesia treatment; a' is the new action, which indicates the anesthetic drug combination and its dosage selected in the new state; max a' Q b,next (s', a') is the maximum expected basic state-action value of all possible actions at the next moment in the new state; Q b,prev (s,a) is the basic state-action value at the previous moment; Q b,next (s', a') is the basic state-action value at the next moment;

[0172] Based on the basic state-action value, nonlinear activation functions and feature weights are introduced to capture nonlinear relationships, threshold parameters and slope parameters are introduced to enhance the adaptability of the deep Q network algorithm, and the discount factors of immediate rewards and future rewards are combined to dynamically adjust parameters through iterative learning to generate optimized state-action values;

[0173] The optimization state-action value is calculated using the following formula:

[0174]

[0175] Among them, Q o (s,a) is the optimized state-action value; s is the current state, which describes the physiological parameters and environmental conditions of the patient during anesthesia; a is the current action, which indicates the anesthetic drug combination and its dosage selected in the current state; α is the learning rate, which determines the balance between the weight of new information and the weight of old information; r is the immediate reward, which indicates the direct benefit or effect brought by the current decision; γ is the discount factor, which is used to measure the importance of future rewards; s' is the new state after the action is executed, which reflects the latest physiological parameters and environmental conditions of the patient after receiving a specific anesthesia treatment; a' is the new action, which indicates the anesthetic drug combination and its dosage selected in the new state; β is the slope parameter of the activation function, which controls the steepness of the nonlinear activation function; w i is the i-th eigenvalue f of the current action a i (a); θ is the threshold parameter, which affects the output range of the activation function; i is the feature index in the anesthetic drug combination feature matrix, from 1 to n; n is the number of features in the anesthetic drug combination feature matrix; f i(a) is the i-th eigenvalue of the current action a; max a' Q o,mext (s', a') is the maximum expected optimal state-action value of all possible actions at the next moment in the new state; Q b (s,a) is the basic state-action value; Q o,next (s′, a′) is the optimized state-action value at the next moment;

[0176] Based on the optimized state-action value, cross-validation technology is used to improve the generalization ability of the optimized state-action value. The optimized state-action value after the generalization ability improvement is used as a prediction variable, and the individualized reaction patterns of different anesthetic drug combinations in the preliminary anesthesia plan are simulated and predicted to generate an individualized reaction prediction pattern.

[0177] This method aims to more accurately capture the individualized response patterns of anesthetic drug combinations for different patients. It introduces a deep Q-network algorithm to dynamically adjust the selection and dosage of anesthetic drugs, and continuously optimizes the anesthesia scheme through the state-action value function update mechanism to ensure its safety and effectiveness. At the same time, it uses nonlinear activation functions and feature weights to enhance the adaptability of the model, and improves the generalization ability through cross-validation to ensure the applicability of the model in different patients.

[0178] Where S is the situation matching score; w i is the weight coefficient of the i-th situation mode; C i is the behavior pattern characteristic of the i-th situation pattern; C ref is the reference behavior pattern characteristic; σ C is the standard deviation of the behavior pattern characteristics; α is the influence coefficient of visual attention distribution; V i is the visual attention distribution in the i-th situation mode; V avg is the average visual attention distribution; V max and V min are the maximum and minimum visual attention distributions, respectively; i is the index of the situation mode, from 1 to N; N is the number of situation modes;

[0179] The present application considers that in order to solve the problem of lack of real-time physiological state monitoring and dynamic adjustment of safety thresholds during anesthesia in the prior art, the present invention embodiment proposes this optional solution to solve the technical problem of personalized anesthesia management, and thus proposes a new optional solution, which includes:

[0180] Based on the potential synergistic effect report, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparison and analysis is performed between the patient's continuous physiological state parameters and the preset safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a real-time comparison and analysis result, including:

[0181] Based on the potential synergy report, the sliding window technology is applied to segment all the time series data in the potential synergy report to capture short-term trends and fluctuation patterns, and the Bayesian inference method is used to estimate the prior distribution of all the time series data to generate the initial safety threshold:

[0182] The initial safety threshold is calculated using the following formula:

[0183]

[0184] Wherein, S0(p,t) is the initial safety threshold of patient p at time t; x(p,t) is the comprehensive index of the continuous physiological state parameter vector of patient p at time t; μ(x(p,t)) is the mean of the continuous physiological state parameter vector, indicating the average physiological state at this time point; σ(x(p,t)) is the standard deviation of the continuous physiological state parameter vector, indicating the degree of fluctuation of the physiological state; z α / 2 is the quantile corresponding to the confidence level, which is used to determine the safety margin of the safety threshold; δ is the adjustment coefficient of the nonlinear term; λ is the slope parameter of the activation function; θ is the threshold parameter, which affects the output range of the activation function;

[0185] Based on the initial safety threshold, a sensitivity parameter is introduced to control the response speed of the nonlinear term, the adaptability of the nonlinear transformation to the abnormal physiological state in the potential synergistic effect report is enhanced, and additional feature weights are introduced to further refine the influence of the patient's continuous physiological state parameters to generate an optimized safety threshold;

[0186] The optimized safety threshold is calculated using the following formula:

[0187]

[0188] Among them, S o (p, t) is the optimized safety threshold; η is the adjustment coefficient, which determines the impact of the nonlinear term on the safety threshold; β' is the sensitivity parameter, which controls the response speed of the nonlinear term; j is the index of the continuous physiological state parameter, from 1 to m; m is the number of continuous physiological state parameters; x j (p, t) is the jth continuous physiological state parameter of patient p at time t; L jis the preset target value of the jth continuous physiological state parameter; γ is the weight coefficient of the additional item; k is the index of the additional feature, from 1 to l; l is the number of additional features; v k g k The weight of (p,t); g k (p,t) is the kth additional eigenvalue of patient p at time t; ρ is the adjustment coefficient of the additional nonlinear term; φ is the slope parameter of the additional nonlinear term; w k g k The nonlinear transformation weight of (p, t); ψ is the threshold parameter of the additional nonlinear term;

[0189] Based on the optimized safety threshold, the real-time parameter adjustment mechanism in the adaptive feedback control algorithm is used to perform a real-time comparative analysis of the patient's continuous physiological state parameters and the optimized safety threshold, and a boundary extension technique is used to expand the optimized safety threshold boundary to generate a real-time comparative analysis result.

[0190] This method aims to improve the safety and effectiveness of the anesthesia process and ensure that the patient maintains stable vital signs during surgery. An adaptive feedback control algorithm is introduced to monitor the patient's continuous physiological state parameters in real time and dynamically adjust them according to the preset safety threshold. The initial safety threshold is estimated through sliding window technology and Bayesian inference method, and the safety threshold is further refined and optimized by combining nonlinear terms and additional feature weights to ensure that the model can quickly respond to abnormal physiological states and maintain high sensitivity and specificity.

[0191] In the initial safety threshold, the mean term μ(x(p,t)): represents the average physiological state at that time point; provides a stable reference baseline; the standard deviation term σ(x(p,t))·z α / 2 : Indicates the degree of fluctuation of physiological state; safety margin used to determine safety threshold; nonlinear adjustment item Enhance the model's adaptability to abnormal physiological states; make the safety threshold more sensitive when the physiological state approaches the critical value;

[0192] Among them, patient p and time t are set according to the actual surgical situation; the continuous physiological state parameter vector x(p,t) is extracted from the real-time monitoring equipment (such as ECG monitor, sphygmomanometer, etc.); the mean μ(x(p,t)) and standard deviation σ(x(p,t)) are calculated by applying statistical software or statistical functions in programming languages ​​to x(p,r); the quantile z corresponding to the confidence level α / 2 According to the required confidence level (such as 95%), the standard normal distribution table is consulted or obtained using statistical software; the adjustment coefficient δ of the nonlinear term is determined by experiments or historical data training model; the slope parameter λ of the activation function is determined by experiments or historical data training model; the threshold parameter θ is determined by experiments or historical data training model;

[0193] In the optimization of the safety threshold, the initial safety threshold term S0(p,t): provides a basic safety threshold reference; the nonlinear adjustment term Control the response speed of nonlinear items and enhance the adaptability to abnormal physiological states; add feature weighting items By assigning different weights to additional features, the influence of continuous physiological state parameters of patients is refined; additional nonlinear terms Improve the model's adaptability to changes in complex physiological states;

[0194] Among them, the initial safety threshold S0(p,t) is calculated by the initial safety threshold formula; the adjustment coefficient η is determined by experiments or historical data training model; the sensitivity parameter β′ is determined by experiments or historical data training model; the continuous physiological state parameter x j (p, t) and the preset target value L j Extracted from real-time monitoring equipment (such as ECG monitors, blood pressure monitors, etc.), the preset target value is set according to clinical guidelines; the weight coefficient γ of the additional item is determined by training the model through experiments or historical data; the additional feature index k and number l are determined according to the number of additional features; the weight v k Learned through the training data set; additional eigenvalue g k (p, t) is extracted from real-time monitoring equipment (such as ECG monitors, sphygmomanometers, etc.); the adjustment coefficient ρ of the additional nonlinear term is determined by experiments or historical data training models; the slope parameter φ of the additional nonlinear term is determined by experiments or historical data training models; the nonlinear transformation weight w k It is learned through the training data set; the threshold parameter ψ of the additional nonlinear term is determined after training the model through experiments or historical data;

[0195] Suppose a female patient is undergoing abdominal tumor resection surgery and her medical record shows a history of hypertension. During anesthesia, the continuous physiological state parameters monitored in real time include heart rate, blood pressure, and respiratory rate; suppose the patient's continuous physiological state parameter vector x(p,t) = [0.7, 0.8, 0.6], mean μ(x(p,t)) = 0.72, standard deviation σ(x(p,t)) = 0.08, and the quantile z corresponding to the confidence level α / 2 =1.96, adjustment coefficient of nonlinear term δ=0.5, slope parameter λ of activation function=0.6, threshold parameter θ=0.7;

[0196]

[0197] Assume that the adjustment coefficient η = 0.5, the sensitivity parameter β' = 0.7, and the continuous physiological state parameter x j (p, t) = [0.7, 0.8, 0.6], preset target value Lj = [0.75, 0.8, 0.65], the weight coefficient of the additional item γ = 0.4, the additional feature index k and number l = 2, the weight v k =[0.3,0.7], additional eigenvalue g k (p, t) = [0.8, 0.9], the adjustment coefficient of the additional nonlinear term ρ = 0.3, the slope parameter of the additional nonlinear term φ = 0.9, the nonlinear transformation weight w k = [0.5, 0.5], the threshold parameter of the additional nonlinear term ψ = 0.4;

[0198]

[0199] Assuming that the threshold is set to 0.9, since the calculated result 0.95 is greater than the set threshold, it shows that the optimized safety threshold has high effectiveness and safety, and can ensure that the patient maintains stable vital signs during surgery. This is because the higher optimized safety threshold reflects that the model can more accurately identify and respond to abnormal physiological states under current conditions, thereby ensuring the safety and effectiveness of the anesthesia process. Through the above steps, the accurate monitoring of the patient's continuous physiological state parameters and the dynamic adjustment of the safety threshold are ensured, which improves the success rate of the operation and the safety of the patient, enhances the accuracy and response speed of the entire anesthesia management system, and realizes personalized anesthesia management.

[0200] Figure 2 A structural diagram of a personalized anesthesia management system is provided for an embodiment of the present application, such as Figure 2 As shown, the device comprises:

[0201] The collection module 21 is used to collect the patient's real-time physiological data stream and previous medical record information, perform a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generate a preliminary anesthesia plan;

[0202] The processing module 22 is used to perform simulation and prediction processing on the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan based on the preliminary anesthesia plan by using a deep Q-network algorithm and a state-action value function update mechanism in the deep Q-network algorithm, and to optimize the anesthetic drug selection and dosage in the preliminary anesthesia plan according to the simulation and prediction processing results by using a dynamic adaptive dose adjustment technology to generate an optimized anesthesia plan;

[0203] The analysis module 23 is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme by using an adaptive feedback control algorithm based on the optimized anesthesia scheme, preset a safety threshold for the patient's continuous physiological state parameters, perform a real-time comparative analysis between the patient's continuous physiological state parameters and the preset safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm, automatically perform a warning process on the trend that deviates from the normal range in the real-time comparative analysis by using an abnormal trend warning technology, and generate a real-time adjustment instruction;

[0204] The generation module 24 is used to obtain the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed based on the real-time adjustment instructions, perform retrospective analysis on the decision points in the process of optimizing the anesthesia plan, and generate a personalized anesthesia management plan.

[0205] Figure 2 The personalized anesthesia management system can perform Figure 1 The implementation principle and technical effect of the personalized anesthesia management method described in the embodiment are not described in detail. The specific way in which each module and unit performs operations in the personalized anesthesia management system in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0206] In one possible design, Figure 2 A personalized anesthesia management system of the illustrated embodiment may be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0207] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0208] The processing component 32 is used to: collect the patient's real-time physiological data stream and previous medical record information, perform a preliminary assessment of the surgical needs based on the real-time physiological data stream and previous medical record information, and generate a preliminary anesthesia plan; use a deep Q network algorithm, through the state-action value function update mechanism in the deep Q network algorithm, simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan, use dynamic adaptive dose adjustment technology, optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation prediction processing results, and generate an optimized anesthesia plan; use an adaptive feedback control algorithm to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, preset a safety threshold for the patient's continuous physiological state parameters, perform real-time comparative analysis of the patient's continuous physiological state parameters and the safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm, use abnormal trend warning technology to automatically warn the trend that deviates from the normal range in the real-time comparative analysis results, and generate real-time adjustment instructions; based on the real-time adjustment instructions, obtain the patient's recovery trajectory and subjective comfort evaluation after the completion of the optimized anesthesia plan, perform retrospective analysis and processing on the decision points in the process of optimizing the anesthesia plan, and generate a personalized anesthesia management plan.

[0209] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0210] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0211] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0212] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0213] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0214] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0215] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A personalized anesthesia management method of the illustrated embodiment.

[0216] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0217] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0218] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0219] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A personalized anesthesia management method, characterized in that: include: Collecting the patient's real-time physiological data stream and previous medical record information, conducting a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generating a preliminary anesthesia plan; Using a deep Q-network algorithm, through the state-action value function update mechanism in the deep Q-network algorithm, simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan, and use dynamic adaptive dose adjustment technology to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation and prediction results to generate an optimized anesthesia plan; Using an adaptive feedback control algorithm, extracting the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, presetting a safety threshold for the patient's continuous physiological state parameters, performing a real-time comparative analysis of the patient's continuous physiological state parameters and the safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm, using an abnormal trend early warning technology, automatically performing early warning processing on the trend of the real-time comparative analysis results that deviates from the normal range, and generating a real-time adjustment instruction; Based on the real-time adjustment instructions, the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed are obtained, and the decision points in the process of optimizing the anesthesia plan are retrospectively analyzed and processed to generate a personalized anesthesia management plan.

2. The method according to claim 1, characterized in that The deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through the state-action value function update mechanism in the deep Q network algorithm, and the dynamic adaptive dose adjustment technology is used to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan according to the simulation prediction processing results to generate an optimized anesthesia plan, including: Extracting anesthetic drugs and dosage information in the preliminary anesthesia plan, classifying and normalizing the anesthetic drugs and dosage information, and generating a standardized anesthetic drug configuration table; Based on the standardized anesthetic drug configuration table, a deep Q-network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through a state-action value function update mechanism in the deep Q-network algorithm, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination; Based on the optimized anesthetic drug combination, a dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing result, so as to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan and generate a drug dose adjustment result; Based on the drug dosage adjustment result, all optimization parameters in the state-action value function optimization process are integrated to generate an optimized anesthesia plan.

3. The method according to claim 2, characterized in that Based on the standardized anesthetic drug configuration table, the deep Q network algorithm is used to simulate and predict the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan through the state-action value function update mechanism in the deep Q network algorithm, and the state-action value function is optimized through iterative learning to generate an optimized anesthetic drug combination, including: By means of a feature extraction method, the pharmacological mechanism features in the standardized anesthetic drug configuration table are extracted, the patient personalized features in the previous medical record information are extracted, the interaction effect between the pharmacological mechanism features and the patient personalized features is calculated, and an anesthetic drug combination feature matrix is ​​generated; Based on the anesthetic drug combination feature matrix, a deep Q-network algorithm is used to simulate and predict the individualized reaction patterns of different anesthetic drug combinations in the preliminary anesthesia plan through a state-action value function update mechanism in the deep Q-network algorithm to generate an individualized reaction prediction pattern; Based on the individualized response prediction model, by introducing a reward mechanism in iterative learning, a positive reward is given to the drug combination that achieves the expected anesthetic effect and has the least side effects among the different anesthetic drug combinations, so as to gradually optimize the state-action value function and generate an optimized state-action value function; The optimal anesthetic drug selection and dosage information in the optimized state-action value function is extracted, and the effect verification processing is performed on the optimal anesthetic drug selection and dosage information to generate an optimized anesthetic drug combination.

4. The method according to claim 2, characterized in that: Based on the optimized anesthetic drug combination, the dynamic adaptive dose adjustment technology is used to quantitatively evaluate the anesthetic effect and side effect risk in the simulation prediction processing result, so as to optimize the anesthetic drug selection and dose in the preliminary anesthesia plan and generate a drug dose adjustment result, including: Normalizing the optimized anesthetic drug combination, quantifying the interactions between the drugs in the optimized anesthetic drug combination, and generating an anesthetic drug combination effect matrix; Using dynamic adaptive dose adjustment technology, the anesthetic drug combination effect matrix is ​​input into the pharmacokinetic model in the dynamic adaptive dose adjustment technology, and the anesthetic effect and side effect risk in the simulation prediction processing result are quantitatively evaluated to generate a quantitative evaluation result; Obtaining an optimization decision basis based on the quantitative evaluation result, optimizing the anesthetic drug selection and dosage in the preliminary anesthesia plan according to the optimization decision basis, and generating an optimized anesthetic drug selection and dosage; Based on the optimized anesthetic drug selection and dosage, the drug combination effect in the anesthetic drug selection and dosage optimization process in the preliminary anesthesia plan is evaluated to generate a drug dosage adjustment result.

5. The method according to claim 1, characterized in that The method uses the adaptive feedback control algorithm to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme, presets a safety threshold for the patient's continuous physiological state parameters, performs real-time comparative analysis on the patient's continuous physiological state parameters and the safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm, uses abnormal trend early warning technology to automatically warn the trend that deviates from the normal range in the real-time comparative analysis results, and generates real-time adjustment instructions, including: Extracting the time-series physiological data during the implementation of the optimized anesthesia scheme, performing time series analysis on the time-series physiological data, recording the patient's physiological dynamic changes during the time series analysis, and generating a dynamic file of the patient's physiological state; Based on the patient's physiological state dynamic file, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparison and analysis is performed between the patient's continuous physiological state parameters and the safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a preset safety threshold range; Based on the preset safety threshold range, an abnormal trend warning technology is used to set up a warning mechanism, and according to the warning mechanism, automatic warning processing is performed on the trend that deviates from the normal range in the real-time comparative analysis to generate an automatic warning result; The abnormal trend in the automatic warning result is recorded, and the anesthesia depth in the optimized anesthesia scheme is adjusted in real time according to the abnormal trend, and a real-time adjustment instruction is generated.

6. The method according to claim 5, characterized in that Based on the patient's physiological state dynamic file, the adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, preset a safety threshold for the patient's continuous physiological state parameters, and perform real-time comparative analysis between the patient's continuous physiological state parameters and the safety threshold through the real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a preset safety threshold range, including: Performing correlation analysis based on the dynamic change parameters in the patient's physiological state dynamic profile, identifying potential synergistic effects in the correlation analysis results, and generating a potential synergistic effect report; Based on the potential synergistic effect report, an adaptive feedback control algorithm is used to extract the patient's continuous physiological state parameters during the implementation of the optimized anesthesia plan, a safety threshold is preset for the patient's continuous physiological state parameters, and a real-time comparative analysis is performed between the patient's continuous physiological state parameters and the preset safety threshold through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm to generate a real-time comparative analysis result; Based on the real-time comparison data in the real-time comparison analysis result and the individual differences of the patients, dynamically adjusting the preset safety threshold to adapt to the fluctuation of the real-time comparison data and the individual differences of the patients, and generating a dynamic safety threshold; The patient's continuous physiological state parameter is compared with the dynamic safety threshold item by item. When the patient's continuous physiological state parameter is greater than the dynamic safety threshold, the upper limit of the dynamic safety threshold is increased to expand the dynamic safety threshold range to generate a preset safety threshold range.

7. The method according to claim 1, characterized in that The method of obtaining the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed based on the real-time adjustment instruction, performing retrospective analysis and processing on the decision points in the process of optimizing the anesthesia plan, and generating a personalized anesthesia management plan includes: Extract all adjustment decision points based on the real-time adjustment instruction, classify all the adjustment decision points, perform logical relationship analysis on different decision point categories in the classification results, and generate a decision point category file; Matching different decision point categories in the decision point category file to the optimized anesthesia plan to clarify the implementation process of the optimized anesthesia plan, obtaining the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed, and generating a recovery trajectory and comfort evaluation report; Performing retrospective analysis on the turning points of the recovery trajectory and the dynamic changes of the comfort evaluation in the recovery trajectory and comfort evaluation report to generate a retrospective analysis report; The improvement suggestions in the retrospective analysis report are summarized, and the improvement suggestions are introduced into the real-time adjustment instructions to optimize the real-time adjustment instructions in a targeted manner and generate a personalized anesthesia management plan.

8. A personalized anesthesia management system, characterized in that: include: A collection module, used to collect the patient's real-time physiological data stream and previous medical record information, conduct a preliminary assessment of surgical needs based on the real-time physiological data stream and previous medical record information, and generate a preliminary anesthesia plan; A processing module is used to perform simulation and prediction processing on the individualized response patterns of different anesthetic drug combinations in the preliminary anesthesia plan based on the preliminary anesthesia plan by using a deep Q-network algorithm and a state-action value function update mechanism in the deep Q-network algorithm, and to optimize the selection and dosage of anesthetic drugs in the preliminary anesthesia plan according to the simulation and prediction processing results by using a dynamic adaptive dose adjustment technology to generate an optimized anesthesia plan; An analysis module, for extracting the patient's continuous physiological state parameters during the implementation of the optimized anesthesia scheme by using an adaptive feedback control algorithm based on the optimized anesthesia scheme, presetting safety thresholds for the patient's continuous physiological state parameters, performing real-time comparative analysis between the patient's continuous physiological state parameters and the preset safety thresholds through a real-time parameter adjustment mechanism in the adaptive feedback control algorithm, automatically performing early warning processing on the trend that deviates from the normal range in the real-time comparative analysis by using an abnormal trend early warning technology, and generating real-time adjustment instructions; A generation module is used to obtain the patient's recovery trajectory and subjective comfort evaluation after the optimized anesthesia plan is completed based on the real-time adjustment instructions, retrospectively analyze and process the decision points in the process of optimizing the anesthesia plan, and generate a personalized anesthesia management plan.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a personalized anesthesia management method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a personalized anesthesia management method as described in any one of claims 1 to 7 is implemented.

Images