Anesthesia postoperative patient state quantitative evaluation method based on multi-feature recognition

Abstract

The invention relates to the technical field of medical data processing, in particular to a multi-feature recognition based postoperative anesthesia patient state quantitative evaluation method. According to the method, the anesthesia postoperative state is divided into a deep anesthesia stage, a superficial consciousness recovery stage and a cognitive recovery stage based on clinical observation. In the deep anesthesia stage, physiological feature data are collected at high frequency, and a physical sign health index is obtained by calculating the standardized deviation between the physiological feature data and a normal reference value; in the shallow consciousness recovery stage, physiological feature data and unconsciousness behavior data are collected, a weighted model is constructed in combination with the physical sign health index, and an unconsciousness behavior response index is obtained; in the cognitive recovery stage, physiological feature data and subjective feature data are collected, and short-term trend prediction is performed on a cognitive recovery index based on a grey prediction model. Furthermore, multi-source feature data of the three stages are fused to construct a dynamic trajectory prediction model, the postoperative recovery trend and state change of the patient are comprehensively analyzed, and quantitative evaluation of the postoperative delirium risk is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a method for quantitatively evaluating the status of a patient after anesthesia using multi-feature recognition. Background Art

[0002] Anesthesia plays an indispensable role in modern surgery, but the complexity of the postoperative awakening process also presents numerous clinical challenges. The postoperative recovery of consciousness typically involves multiple stages, with significant differences in neurological activity, physiological parameters, and behavioral manifestations at different stages. Failure to monitor or accurately assess postoperative consciousness can lead to serious complications such as postoperative delirium and cognitive impairment.

[0003] Currently, clinical assessment of post-anesthesia status primarily relies on medical staff's observation of vital signs and subjective behavioral manifestations. This lacks a systematic, quantitative assessment method, and the results are highly subjective and lagging. Some studies have attempted to use a single physiological indicator to assess status, but due to significant individual variability and the complex nature of postoperative status, a single indicator alone cannot fully reflect a patient's true recovery status.

[0004] Furthermore, traditional prediction methods, often using models like linear regression and logistic regression, struggle to adapt to practical challenges such as limited data samples, unclear trends, and weak cross-stage feature correlations in postoperative status data. Effective prediction mechanisms are still lacking for the early identification of high-risk postoperative complications, such as delirium.

[0005] To this end, the present invention proposes a multi-feature recognition method for quantitatively evaluating the status of patients after anesthesia. Summary of the Invention

[0006] The present invention provides a multi-feature recognition method for quantitatively assessing the status of patients after anesthesia, aiming to address the problems of existing postoperative status assessment methods, which are highly subjective, poor real-time performance, and insufficient predictive ability. This method divides the postoperative status into the deep anesthesia stage, the shallow consciousness recovery stage, and the cognitive recovery stage through clinical observation. Physiological characteristics, unconscious behavioral characteristics, and subjective characteristic data are collected separately. A physical health index is calculated based on standardized deviations, and a weighted model is constructed to obtain a behavioral response index. A gray prediction model is used to predict the short-term trend of the cognitive recovery index. Furthermore, a dynamic trajectory prediction model is established based on data from each stage. This model dynamically analyzes the patient's postoperative recovery trajectory, enabling quantitative assessment and early warning of complications such as postoperative delirium.

[0007] To achieve the above object, the present invention provides the following technical solutions: A multi-feature recognition method for quantitatively evaluating the status of patients after anesthesia surgery, comprising: Based on clinical observation of patients, the postoperative state of anesthesia is divided into deep anesthesia stage, shallow consciousness recovery stage and cognitive recovery stage; During the deep anesthesia stage, physiological characteristic data are collected at a first frequency and uploaded to a database, and the physical sign health index is obtained by calculating the standardized deviation between the physiological characteristic data and the normal reference value; During the shallow consciousness recovery phase, physiological characteristic data and unconscious behavior data are collected at the second frequency and uploaded to the database. A weighted model is constructed based on the physical sign health index to obtain the unconscious behavior response index. In the cognitive recovery stage, physiological and subjective characteristic data were collected at the third frequency and uploaded to the database, and the grey prediction model was used to make short-term predictions of the cognitive recovery index. Based on the physiological characteristic data, unconscious behavior data and subjective characteristic data obtained at each stage, a dynamic trajectory prediction model was established, and a comprehensive analysis was performed in combination with the short-term prediction results of the grey prediction model to achieve a quantitative assessment of the risk of postoperative delirium.

[0008] Furthermore, the first frequency is greater than the second frequency, and the second frequency is greater than the third frequency.

[0009] Furthermore, the steps of obtaining the physical health index include: Collect physiological characteristic data, calculate the standardized difference between each physiological characteristic data and the preset normal reference value, and obtain the standardized deviation; According to the standardized deviation of each physiological characteristic data, the physical sign health index is obtained through the physical sign health calculation formula; The calculation formula of the physical sign health index is: ; in, Indicates the health index of physical signs during the deep anesthesia stage, Indicates the total number of monitored physiological characteristics, Indicates the The measured values ​​of physiological characteristics, Indicates the Normalized value of a physiological characteristic.

[0010] Furthermore, the steps of obtaining the unconscious behavior response index include: Collect behavioral characteristic data related to the patient's unconscious state, and perform weighted fusion of the behavioral characteristic data with the synchronously collected physical health index; The behavioral response score is calculated according to the set weighted fusion formula as the unconscious behavioral response index; The calculation formula of the weighted fusion formula is: ; in, represents the unconscious behavioral response index, Indicates the physical health index of the shallow consciousness recovery stage, Indicates the total number of unconscious behaviors monitored, Indicates the Normalized value of unconscious behavior, Represents the weighting coefficient.

[0011] Furthermore, the steps of short-term prediction of cognitive recovery index include: The physiological characteristic data and subjective characteristic data collected during the cognitive recovery phase are used as input vectors; The input vector is accumulated and generated to construct the grey prediction model GM(1,N); Based on the constructed grey prediction model, the short-term trend prediction of cognitive recovery index was carried out.

[0012] Furthermore, the steps of establishing a dynamic trajectory prediction model include: Collect physiological characteristic data, unconscious behavioral data, and subjective characteristic data during the deep anesthesia stage, shallow consciousness recovery stage, and cognitive recovery stage, and perform time series processing; Standardize and extract features from data at each stage to form a stage-by-stage state vector; The sequence modeling algorithm is used to construct a time correlation model and establish a dynamic trajectory prediction model covering all stages.

[0013] Furthermore, the steps of conducting comprehensive analysis based on the short-term forecast results of the grey forecast model include: Matching analysis was performed between the cognitive recovery index output by the grey prediction model and the recovery trajectory output by the dynamic trajectory prediction model; Calculate the deviation trend and consistency level between the two; If the deviation exceeds the set range or the trajectory trend is abnormal, it is determined that there is a high risk of postoperative delirium, an assessment report is generated, and an early warning signal is triggered.

[0014] Furthermore, at any stage, if the normalized deviation of a certain physiological characteristic data is greater than the deviation threshold, it is determined that the physiological state is abnormal and an alarm is issued.

[0015] The beneficial effects of the present invention are: 1. Using a gray prediction model to predict the short-term cognitive recovery index, this method can provide timely predictions of the numerical changes in a patient's postoperative cognitive status, even when data samples are limited and the trend of change is unclear. This method uses physiological characteristic data and subjective response data as input, constructs a GM(1,N) gray model, and dynamically outputs the numerical trend of the cognitive recovery index over a period of time. This helps medical staff quickly identify abnormal recovery or stagnation risks in the early stages of cognitive recovery, enabling proactive intervention for postoperative delirium.

[0016] 2. By establishing a dynamic trajectory prediction model covering all postoperative stages, integrating physiological, behavioral, and subjective data from deep anesthesia, shallow consciousness recovery, and cognitive recovery, a time series modeling of the entire patient recovery process is achieved. This model can predict the patient's overall recovery trend and identify potential nonlinear abnormal changes in the trajectory in advance. When abnormal fluctuations in physiological parameters or behavioral response curves are identified, the system can automatically indicate risks, thereby assisting medical staff in continuous monitoring and timely intervention, improving postoperative safety and management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 The present invention provides a flowchart of a method for quantitatively evaluating the status of a patient after anesthesia using multi-feature recognition. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0019] A multi-feature recognition method for quantitatively evaluating the status of patients after anesthesia surgery, comprising: S100: Based on clinical observation of patients, the postoperative state of anesthesia is divided into deep anesthesia stage, shallow consciousness recovery stage and cognitive recovery stage; Specifically, based on existing clinical practice, a set of postoperative recovery-related indicators that can be monitored in real time or periodically assessed will be selected as the basis for determining status staging. These indicators will include at least: consciousness assessment indicators (using the Glasgow Coma Scale (GCS) or the Rating of Serenity Scale (RASS); neurological reflexes or stimulus responses (such as the presence or absence of responses to sound, pain, and light); autonomic nervous function indicators (including pupillary reflexes, heart rate variability, and respiratory rhythm); and physiological stability indicators (such as blood pressure, body temperature, and blood oxygen saturation).

[0020] The stage division criteria are set based on the consciousness state assessment indicators, including: Deep anesthesia stage (Phase I): The judgment criteria are that the patient is in a state of unconsciousness, has no response to external stimuli, a GCS score ≤ 6, or a RASS score between -5 and -4. During this stage, physiological parameters such as heart rate and blood pressure fluctuate little and are in a stable state under anesthesia maintenance.

[0021] Phase II: The patient's response to mild stimulation (such as tapping the shoulder or voice calling) is slow or unconscious, with a GCS score between 7 and 11 or a RASS score between -3 and -1. This stage may be accompanied by slight involuntary limb movements.

[0022] Cognitive recovery stage (Phase III): The judgment criteria are that the patient has the ability to understand and respond to verbal instructions, a GCS score ≥ 12, or a RASS score above 0. In this stage, the subjective consciousness expression ability (such as speaking and cooperating with instructions) gradually recovers.

[0023] After dividing the post-anesthesia status according to existing clinical practice data, various patient data were subsequently collected at the first frequency, second frequency, and third frequency.

[0024] Furthermore, the first frequency is greater than the second frequency, and the second frequency is greater than the third frequency.

[0025] Specifically, during the deep anesthesia stage, the sampling frequency of physiological characteristics is set to continuous collection in seconds (such as once every 5 seconds) to ensure high-frequency monitoring of changes in vital signs in the early stage after anesthesia; during the shallow consciousness recovery stage, the sampling frequency of physiological and behavioral data is moderately reduced and can be set to once every 5 to 10 minutes; during the cognitive recovery stage, the physiological state changes relatively slowly, and the frequency is further reduced, which can be set to a subjective state assessment every 30 minutes to 1 hour.

[0026] During deep anesthesia, patients haven't yet regained consciousness and are highly dependent on drug metabolism and fluctuations in vital signs. Any abnormalities in physiological parameters such as heart rate, respiration, and blood pressure can have serious consequences, necessitating high sampling rates and real-time responsiveness. During the shallow consciousness recovery phase, patients begin to exhibit unconscious reflexes or behaviors, but overall physiological fluctuations stabilize. At this point, the sampling frequency of physiological features can be appropriately reduced, and recognizable non-command behavioral data (such as eye opening, grasping, and limb movement) can be collected. The sampling frequency should be appropriately adjusted based on the intermittent nature of the behavior. During the cognitive recovery phase, patients gradually regain cognitive abilities and enhance their subjective interaction abilities, but their physiological state remains relatively stable. Subjective features (such as simple conversations, pain ratings, and command execution) are less frequency-sensitive, and excessive sampling can increase interference. Therefore, a lower sampling frequency is employed.

[0027] S200: During the deep anesthesia stage, physiological characteristic data is collected at a first frequency and uploaded to a database, and a physical sign health index is obtained by calculating a standardized deviation between the physiological characteristic data and a normal reference value; Furthermore, the steps of obtaining the physical health index include: Collect physiological characteristic data, calculate the standardized difference between each physiological characteristic data and the preset normal reference value, and obtain the standardized deviation; According to the standardized deviation of each physiological characteristic data, the physical sign health index is obtained through the physical sign health calculation formula; The calculation formula of the physical sign health index is: ; in, Indicates the health index of physical signs during the deep anesthesia stage, Indicates the total number of monitored physiological characteristics, Indicates the The measured values ​​of physiological characteristics, Indicates the Normalized value of a physiological characteristic.

[0028] Specifically, during the deep anesthesia stage, the patient's multiple key physiological characteristic parameters are collected at a high frequency by monitoring equipment, including but not limited to heart rate, respiratory rate, systolic blood pressure, diastolic blood pressure, blood oxygen saturation and body temperature. , compare the measured value with the preset normal reference value Compare and compare by standard deviation Perform standardization and calculate standardized deviation , the standardization process can be expressed as: Based on the standardized deviation, the physical health index Perform a comprehensive evaluation. If the deviation of all characteristics is small, then If it is close to 1, it means that the patient's physiological state is close to normal; if some characteristics deviate greatly, then Decreased, suggesting that the patient has potential physiological risks.

[0029] By constructing a physical health index, which uses standardized deviations to comprehensively assess multiple physiological characteristics and form a unified numerical indicator, we can help accurately quantify the overall physiological stability of patients during deep anesthesia. This index reflects the degree to which an individual deviates from the normal physiological range, providing a basis for subsequent behavioral feature weighting and status prediction. This proposal aims to disclose a reasonable evaluation standard and provide a reference value for postoperative monitoring.

[0030] S300: In the shallow consciousness recovery stage, physiological characteristic data and unconscious behavior data are collected at the second frequency and uploaded to the database, and a weighted model is constructed based on the physical sign health index to obtain the unconscious behavior response index; Furthermore, the steps of obtaining the unconscious behavior response index include: Collect behavioral characteristic data related to the patient's unconscious state, and perform weighted fusion of the behavioral characteristic data with the synchronously collected physical health index; The behavioral response score is calculated according to the set weighted fusion formula as the unconscious behavioral response index; The calculation formula of the weighted fusion formula is: ; in, represents the unconscious behavioral response index, Indicates the physical health index of the shallow consciousness recovery stage, Indicates the total number of unconscious behaviors monitored, Indicates the Normalized value of unconscious behavior, Represents the weighting coefficient.

[0031] Specifically, during the shallow consciousness recovery stage, the patient gradually develops involuntary neuromuscular reactions or non-verbal behaviors. Wearable devices, camera systems, or bedside monitoring devices are used to collect the following unconscious behavior data: eye opening reaction frequency, limb twitching or movement, blinking frequency, grasping reflex, head rotation, responsiveness, etc. During this stage, physiological characteristic data continues to be collected at a second frequency to synchronize the calculation of the vital sign health index at this stage. . Similarly, the original observation value of each unconscious behavior Normalize it to get its dimensionless representation In order to reflect the importance of each behavioral characteristic and combine it with the current physiological state, the behavioral characteristic data is weighted and fused with the synchronously collected physical health index. The calculation formula of the weighted fusion formula is: ; Among them, the weighting coefficient Through deep anesthesia The average value of can be expressed as ,in Indicates the number of data acquisitions during the deep anesthesia stage.

[0032] By constructing the unconscious behavior response index, integrating the patient's current physical health index with multiple unconscious behavior characteristics, a quantitative assessment of the patient's status in the shallow consciousness recovery stage is achieved. Coordinate the influence weight between physiological stability and behavioral performance, and the patient's The higher the value, the more stable the characteristic state is, so the adaptability reduces the weight of this part, and vice versa. This scheme aims to disclose a reasonable evaluation standard to provide a reference value for postoperative monitoring.

[0033] S400: In the cognitive recovery stage, physiological characteristic data and subjective characteristic data are collected at the third frequency and uploaded to the database, and a grey prediction model is used to perform a short-term prediction of the cognitive recovery index; Furthermore, the steps of short-term prediction of cognitive recovery index include: The physiological characteristic data and subjective characteristic data collected during the cognitive recovery phase are used as input vectors; The input vector is accumulated and generated to construct the grey prediction model GM(1,N); Based on the constructed grey prediction model, the short-term trend prediction of cognitive recovery index was carried out.

[0034] Specifically, during the cognitive recovery phase, the following two types of data are collected at the third frequency, including physiological characteristic data: such as heart rate, respiratory rate, systolic blood pressure, diastolic blood pressure, blood oxygen saturation, and body temperature; and subjective characteristic data, such as the Mini-Mental State Examination scale score or the postoperative question-answering ability test score. These features are combined into a multidimensional input vector in time series. , where the last item is subjective feature data and the rest are physiological feature data. Perform an accumulation generation operation, namely: ; in, represents the original observation data, Represents the data sequence after cumulative generation. According to the data sequence after cumulative generation, the GM(1,N) grey model is constructed, and its mathematical form is: ; in, represents the cumulative series of cognitive recovery index, For the The cumulative sequence of influencing factors; and The parameter to be estimated is obtained by the least squares method. By solving the GM(1,N) differential equation, the short-term predicted value of the cognitive recovery index is obtained, and the de-accumulation process is performed to restore the original value to obtain the numerical sequence of the cognitive recovery index during the prediction period. This part can be expressed as: .

[0035] By constructing a gray prediction model for the cognitive recovery index, we achieved short-term quantitative prediction of cognitive recovery trends, helping clinicians promptly identify potential postoperative delirium risks. The GM(1,N) model, which integrates multidimensional physiological and subjective data, is suitable for practical scenarios with limited postoperative sample size and high data volatility, improving the stability and adaptability of the assessment.

[0036] S500: Based on the physiological characteristic data, unconscious behavior data and subjective characteristic data obtained at each stage, a dynamic trajectory prediction model is established, and a comprehensive analysis is performed in combination with the short-term prediction results of the grey prediction model to achieve a quantitative assessment of the risk of postoperative delirium.

[0037] Furthermore, the steps of establishing a dynamic trajectory prediction model include: Collect physiological characteristic data, unconscious behavioral data, and subjective characteristic data during the deep anesthesia stage, shallow consciousness recovery stage, and cognitive recovery stage, and perform time series processing; Standardize and extract features from data at each stage to form a stage-by-stage state vector; The sequence modeling algorithm is used to construct a time correlation model and establish a dynamic trajectory prediction model covering all stages.

[0038] Specifically, physiological characteristic data, unconscious behavioral data, and subjective characteristic data are collected from patients throughout their postoperative course at pre-set time points during the deep anesthesia phase, the shallow consciousness recovery phase, and the cognitive recovery phase. All data are sorted by acquisition time and processed uniformly into time series to construct a continuous time series input stream.

[0039] The collected multimodal data is standardized to keep different data dimensions within the same numerical range. Statistical feature extraction (such as mean, variance, and frequency domain features) is then used to extract representative features for each stage. The extracted features for each stage are combined into a stage-by-stage state vector, reflecting the patient's overall condition at each stage.

[0040] A sequence modeling algorithm is used to model the sequence of staged state vectors. This algorithm can be a long short-term memory network, a gated recurrent unit, or a Transformer model, which is not limited in this embodiment. The model learns how the patient's state changes over time, establishes a time-dependent model covering all stages, and generates a series of chronologically ordered predictions for the patient's postoperative recovery process. These predictions constitute a dynamic trajectory prediction curve.

[0041] By integrating multi-source data from the deep anesthesia phase, the shallow consciousness recovery phase, and the cognitive recovery phase, a time-dependent model covering the entire cycle is constructed to continuously track and predict the changing trends of patients' postoperative status. This model can reflect the evolution of individual status over time, providing data support for the early identification of abnormal recovery trajectories and improving early warning capabilities for complications such as postoperative delirium.

[0042] Furthermore, the steps of conducting comprehensive analysis based on the short-term forecast results of the grey forecast model include: Matching analysis was performed between the cognitive recovery index output by the grey prediction model and the recovery trajectory output by the dynamic trajectory prediction model; Calculate the deviation trend and consistency level between the two; If the deviation exceeds the set range or the trajectory trend is abnormal, it is determined that there is a high risk of postoperative delirium, an assessment report is generated, and an early warning signal is triggered.

[0043] Specifically, the sequence of cognitive recovery index prediction values ​​generated by the gray prediction model during the cognitive recovery phase serves as the first set of prediction data, while the sequence of state prediction values ​​covering all phases, generated by the dynamic trajectory prediction model, serves as the second set of prediction data. These two sets of prediction data are then aligned along the temporal dimension, and their deviation values ​​(such as mean squared error, mean absolute error, etc.) and trend consistency indicators (such as the Pearson correlation coefficient, dynamic time warping (DTW) score, or directional consistency) are calculated to measure the similarity and degree of deviation in the prediction trends of the two models. A deviation threshold and a trend deviation tolerance interval are set as the basis for determining the risk of postoperative delirium. If there is a significant deviation between the gray model and the trajectory prediction model (such as a deviation value exceeding a threshold or a predicted trend in opposite directions), the patient is considered to have abnormal cognitive recovery and a high risk of postoperative delirium. An assessment report is generated and a warning signal is triggered.

[0044] The cognitive recovery index output by the gray prediction model provides quantitative predictions at the short-term numerical level, while the recovery curve output by the dynamic trajectory prediction model reflects the changing trend of the patient's status in the time series. The two complement each other and are analyzed synergistically. By calculating the deviation trend and consistency level between the two, accurate identification of abnormal recovery patterns can be achieved, which can effectively improve the sensitivity and accuracy of postoperative delirium risk assessment. In addition, when there is a significant difference or trend divergence between the results of the two models, it can be used as a risk signal for atypical conditions, helping to trigger intervention mechanisms in a timely manner to prevent delayed diagnosis.

[0045] Furthermore, at any stage, if the normalized deviation of a certain physiological characteristic data is greater than the deviation threshold, it is determined that the physiological state is abnormal and an alarm is issued.

[0046] When the system detects that the fluctuation amplitude of physiological characteristic indicators exceeds the preset threshold within a certain period of time, indicating that the patient is in a potential physiological unstable state, it can immediately trigger the intervention mechanism or clinical review to achieve intervention and early response at any time.

[0047] Example 2 This embodiment provides a multi-feature recognition method for quantitatively assessing the status of patients after anesthesia surgery, which is used to evaluate the patient's consciousness and cognitive status in stages throughout the postoperative recovery process and to quantitatively analyze the risk of postoperative delirium.

[0048] Based on postoperative bedside clinical observations (such as Glasgow Coma Scale, RASS agitation score, pupillary reaction, and pain response), the patient's postoperative status is divided into the following three stages: Deep anesthesia stage: no autonomous consciousness, slow reflexes, and need to rely on instrument monitoring.

[0049] Shallow consciousness recovery stage: Involuntary behavioral reactions appear, such as opening eyes, moving limbs, changes in breathing rhythm, etc.

[0050] Cognitive recovery stage: The patient can provide meaningful feedback, communicate subjectively or cooperate with simple instructions.

[0051] Physiological characteristic data such as heart rate, blood pressure, respiratory rate, brain electrical index (BIS), and blood oxygen saturation are collected at the first frequency (every 10 seconds) and uploaded to the database.

[0052] The physical sign health index is obtained through the physical sign health calculation formula, where the calculation formula of the physical sign health index is: ; in, Indicates the health index of physical signs during the deep anesthesia stage, Indicates the total number of monitored physiological characteristics, Indicates the The measured values ​​of physiological characteristics, Indicates the Normalized value of a physiological characteristic.

[0053] Physiological characteristic data and unconscious behavioral data (such as voluntary eye opening, electromyography, limb micro-movements, sound response, etc.) are collected at a second frequency (for example, every 10 minutes).

[0054] The following weighted fusion model is used to obtain the unconscious behavioral response index (BRI): ; Physiological characteristics (such as heart rate variability, EEG rhythm) and subjective characteristics (such as patient speech, cooperation, and simple cognitive test scores) are collected at a third frequency (for example, every 60 minutes).

[0055] The data from this phase was constructed as an input vector, and the GM(1,N) grey prediction model was applied to predict the short-term trend of the cognitive recovery index. The prediction results were used to determine the speed and risk of cognitive function recovery.

[0056] All physiological, behavioral, and subjective characteristics of the three phases were organized into time series and standardized to construct a phase-specific state vector. A dynamic trajectory prediction model covering all phases was established using time series modeling algorithms such as LSTM. Combining trajectory prediction trends with short-term numerical predictions from a gray model, the recovery path was analyzed to determine whether it deviated from the normal trajectory and to determine the risk level of postoperative delirium.

[0057] Finally, it should be noted that the above content is merely an example and explanation of the structure of the present invention. Technicians in this technical field may make various modifications or additions to the specific embodiments described, or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should fall within the scope of protection of the present invention.

[0058] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0062] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0063] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition, characterized in that: include: Based on clinical observation of patients, the postoperative state of anesthesia is divided into deep anesthesia stage, shallow consciousness recovery stage and cognitive recovery stage; During the deep anesthesia stage, physiological characteristic data are collected at a first frequency and uploaded to a database, and the physical sign health index is obtained by calculating the standardized deviation between the physiological characteristic data and the normal reference value; During the shallow consciousness recovery phase, physiological characteristic data and unconscious behavior data are collected at the second frequency and uploaded to the database. A weighted model is constructed based on the physical sign health index to obtain the unconscious behavior response index. In the cognitive recovery stage, physiological and subjective characteristic data were collected at the third frequency and uploaded to the database, and the grey prediction model was used to make short-term predictions of the cognitive recovery index. Based on the physiological characteristic data, unconscious behavior data and subjective characteristic data obtained at each stage, a dynamic trajectory prediction model was established, and a comprehensive analysis was performed in combination with the short-term prediction results of the grey prediction model to achieve a quantitative assessment of the risk of postoperative delirium.

2. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The first frequency is greater than the second frequency, and the second frequency is greater than the third frequency.

3. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The steps to obtain a physical health index include: Collect physiological characteristic data, calculate the standardized difference between each physiological characteristic data and the preset normal reference value, and obtain the standardized deviation; According to the standardized deviation of each physiological characteristic data, the physical sign health index is obtained through the physical sign health calculation formula; The calculation formula of the physical sign health index is: ； in, Indicates the health index of physical signs during the deep anesthesia stage, Indicates the total number of monitored physiological characteristics, Indicates the The measured values ​​of physiological characteristics, Indicates the Normalized value of a physiological characteristic.

4. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The steps to obtain the automatic behavioral response index include: Collect behavioral characteristic data related to the patient's unconscious state, and perform weighted fusion of the behavioral characteristic data with the synchronously collected physical health index; The behavioral response score is calculated according to the set weighted fusion formula as the unconscious behavioral response index; The calculation formula of the weighted fusion formula is: ； in, represents the unconscious behavioral response index, Indicates the physical health index of the shallow consciousness recovery stage, Indicates the total number of unconscious behaviors monitored, Indicates the Normalized value of unconscious behavior, Represents the weighting coefficient.

5. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The steps for short-term prediction of cognitive recovery index include: The physiological characteristic data and subjective characteristic data collected during the cognitive recovery phase are used as input vectors; The input vector is accumulated and generated to construct the grey prediction model GM(1,N); Based on the constructed grey prediction model, the short-term trend prediction of cognitive recovery index was carried out.

6. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The steps to build a dynamic trajectory prediction model include: Collect physiological characteristic data, unconscious behavioral data, and subjective characteristic data during the deep anesthesia stage, shallow consciousness recovery stage, and cognitive recovery stage, and perform time series processing; Standardize and extract features from data at each stage to form a stage-by-stage state vector; The sequence modeling algorithm is used to construct a time correlation model and establish a dynamic trajectory prediction model covering all stages.

7. The method for quantitatively evaluating the status of patients after anesthesia based on multi-feature recognition according to claim 1, characterized in that: The steps of conducting comprehensive analysis based on the short-term forecast results of the grey forecast model include: Matching analysis was performed between the cognitive recovery index output by the grey prediction model and the recovery trajectory output by the dynamic trajectory prediction model; Calculate the deviation trend and consistency level between the two; If the deviation exceeds the set range or the trajectory trend is abnormal, it is determined that there is a high risk of postoperative delirium, an assessment report is generated, and an early warning signal is triggered.

8. The method for quantitatively evaluating the status of a patient after anesthesia based on multi-feature recognition according to claim 1, characterized in that: At any stage, if the standardized deviation of a certain physiological characteristic data is greater than the deviation threshold, it is determined to be an abnormal physiological state and an alarm is issued.

Images