Postoperative recovery time determination method, device, equipment, medium and program product

Abstract

The invention discloses a postoperative recovery time determination method and device, equipment, a medium and a program product, and is applied to the technical field of artificial intelligence. The method comprises the following steps: acquiring user parameters of a user; the user parameters comprise physiological feature parameters used for representing the physiological state of the user, disease risk level parameters used for representing the disease of the user, operation related parameters, anesthesia related parameters and medical team ability evaluation parameters; the user parameters are input into an intraoperative anesthesia mode determining model, an intraoperative anesthesia mode output by the intraoperative anesthesia mode determining model is obtained, and the intraoperative anesthesia mode comprises a non-general anesthesia mode or a general anesthesia mode; according to the intraoperative anesthesia mode and the user parameters, the postoperative recovery time of the user is determined. According to the method, the postoperative recovery time of the user is determined through the user parameters of the user in multiple dimensions and the intraoperative anesthesia mode, and the problem that the accuracy is low when the postoperative recovery time of the user is predicted only through the physiological feature data in a single dimension can be solved.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and in particular relates to a method, device, equipment, medium and program product for determining postoperative recovery time. Background Technology

[0002] Cardiovascular interventional surgery is a minimally invasive technique for treating heart and vascular diseases. Its core characteristic is the use of catheters and other interventional devices to directly access the lesion site via a vascular pathway, allowing diagnosis or treatment to be completed without open-chest surgery. Within the technical system of cardiovascular interventional surgery, transcatheter aortic valve replacement (TAVR) is a minimally invasive interventional technique for treating aortic stenosis.

[0003] In the postoperative management of TAVR patients, rehabilitation plans can be developed by predicting postoperative recovery time. Related technologies can assess postoperative recovery time based on patients' physiological characteristics, which may include heart rate, blood pressure, blood oxygen saturation, routine and biochemical indicators, or myocardial injury markers. However, the accuracy of assessing postoperative recovery time based solely on a single dimension of physiological data is low. Summary of the Invention

[0004] This application provides a method, apparatus, device, medium, and program product for determining postoperative recovery time, which can solve the problem of low accuracy in predicting the postoperative recovery time of TAVR patients based solely on single-dimensional physiological characteristic data.

[0005] In a first aspect, embodiments of this application provide a method for determining postoperative recovery time. This method includes: acquiring user parameters; the user parameters include physiological characteristic parameters characterizing the user's physiological state, disease risk level parameters characterizing the user's disease, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters; inputting the user parameters into an intraoperative anesthesia method determination model to obtain the intraoperative anesthesia method output by the model, wherein the intraoperative anesthesia method includes non-general anesthesia or general anesthesia; and determining the user's postoperative recovery time based on the intraoperative anesthesia method and the user parameters.

[0006] Secondly, embodiments of this application provide a postoperative recovery time determination device, which includes: an acquisition module for acquiring user parameters; the user parameters include physiological characteristic parameters characterizing the user's physiological state, disease risk level parameters characterizing the user's disease, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters; a first determination module for inputting the user parameters into an intraoperative anesthesia method determination model to obtain the intraoperative anesthesia method output by the intraoperative anesthesia method determination model, the intraoperative anesthesia method including non-general anesthesia or general anesthesia; and a second determination module for determining the user's postoperative recovery time based on the intraoperative anesthesia method and the user parameters.

[0007] Thirdly, embodiments of this application provide an electronic device, the device comprising: a processor and a memory storing computer program instructions; the processor, when executing the computer program instructions, implements the postoperative recovery time determination method as described in any of the first aspects.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the postoperative recovery time determination method as described in any of the first aspects.

[0009] Fifthly, embodiments of this application provide a computer program product, which includes a computer program or instructions, and when the computer program or instructions are executed by a processor, implements the postoperative recovery time determination method as described in any of the first aspects.

[0010] The postoperative recovery time determination method, apparatus, device, medium, and program product of this application can acquire multi-dimensional user parameters, specifically including physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters. These multi-dimensional user parameters comprehensively depict the factors affecting the user's postoperative recovery from multiple perspectives, such as the user's own physiological state, disease risk, surgical operation, anesthesia effects, and the medical team's level, avoiding the limitations of single-dimensional data. Through the intraoperative anesthesia method determination model, the specific intraoperative anesthesia method is first determined based on the multi-dimensional user parameters, and then the postoperative recovery time is determined by combining the multi-dimensional user parameters and the intraoperative anesthesia method. This makes the postoperative recovery time determination process more closely reflect the user's actual clinical scenario. In this way, more key variables affecting the user's postoperative recovery are integrated, effectively improving the accuracy of the determined postoperative recovery time. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a method for determining postoperative recovery time provided in some embodiments of this application is shown;

[0013] Figure 2 A flowchart illustrating a specific implementation of step 120 provided in some embodiments of this application is shown;

[0014] Figure 3 The diagram illustrates a specific implementation of step 1202 provided in some embodiments of this application;

[0015] Figure 4 The diagram illustrates a specific implementation of step 130 provided in some embodiments of this application;

[0016] Figure 5 A flowchart illustrating a specific implementation of step 1303 provided in some embodiments of this application is shown;

[0017] Figure 6 A flowchart illustrating the method for determining a target recovery time in some embodiments of this application is shown.

[0018] Figure 7 The diagram illustrates a flowchart of the method for generating postoperative management strategies in the postoperative recovery time determination method provided in some embodiments of this application;

[0019] Figure 8 The present application provides a schematic diagram of the structure of a postoperative recovery time determination device according to some embodiments;

[0020] Figure 9 The diagram shows a schematic representation of the structure of an electronic device provided in some embodiments of this application. Detailed Implementation

[0021] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0023] It should be noted that the acquisition, storage, use, and processing of data in this application embodiment all comply with the relevant provisions of national laws and regulations.

[0024] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0025] Before describing the technical solutions provided in the embodiments of this application, in order to facilitate understanding of the embodiments of this application, this application first provides a detailed description of the relevant technologies involved:

[0026] Transcatheter aortic valve replacement (TAVR), as a minimally invasive interventional technique for treating aortic stenosis, has become the preferred treatment option for high-risk and very high-risk patients due to its advantages of minimal trauma and rapid recovery. With the continuous development of TAVR and the gradual expansion of its indications, the number of TAVR procedures has been increasing year by year. Against this backdrop, the need for accurate prediction of postoperative recovery time for TAVR patients is becoming increasingly urgent. Understandably, accurate prediction of postoperative recovery time not only helps optimize the allocation of medical resources and develop personalized rehabilitation plans, but also improves patient satisfaction and the quality of medical care.

[0027] In the perioperative management of TAVR, the choice of anesthesia method affects the patient's postoperative recovery time. Currently, the main clinical methods used are general anesthesia (GA) and non-general anesthesia. Non-general anesthesia includes monitored anesthesia care (MAC) (i.e., local anesthesia with sedation (LA+S)) and local anesthesia alone (LA). Studies have shown that different anesthesia methods differ in terms of awakening time, cognitive recovery, and recovery of motor function, thus affecting the patient's overall recovery process.

[0028] In related technologies, when assessing the postoperative recovery time of TAVR patients, postoperative recovery time can be determined based on the patient's physiological characteristics, such as heart rate, blood pressure, blood oxygen saturation, routine and biochemical indicators, or myocardial injury markers. Therefore, determining the postoperative recovery time of TAVR patients based solely on a single dimension of physiological characteristic data has low accuracy.

[0029] To address the problems in the aforementioned related technologies, embodiments of this application provide a method, apparatus, device, medium, and program product for determining postoperative recovery time. The following is a detailed description in conjunction with the appendix. Figure 1 To be continued Figure 6 The method for determining postoperative recovery time provided in this application will be described in detail through specific embodiments and application scenarios.

[0030] Figure 1 A flowchart illustrating a method for determining postoperative recovery time provided in some embodiments of this application is shown. For example... Figure 1 As shown, the method for determining postoperative recovery time may include steps 110 to 130.

[0031] Step 110: Obtain the user's parameters; these parameters include physiological characteristic parameters characterizing the user's physiological state, disease risk level parameters characterizing the user's disease, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters; Step 120: Input the user parameters into the intraoperative anesthesia method determination model to obtain the intraoperative anesthesia method output by the model, which includes non-general anesthesia or general anesthesia; Step 130: Determine the user's postoperative recovery time based on the intraoperative anesthesia method and user parameters.

[0032] This allows for the acquisition of multi-dimensional user parameters, including physiological characteristics, disease risk level, surgery-related parameters, anesthesia-related parameters, and medical team competence evaluation parameters. These multi-dimensional user parameters comprehensively characterize factors influencing postoperative recovery from multiple perspectives, such as the user's physiological state, disease risk, surgical procedure, anesthesia effects, and the medical team's skill level, avoiding the limitations of single-dimensional data. The intraoperative anesthesia method determination model first determines the specific intraoperative anesthesia method based on the multi-dimensional user parameters, and then combines these parameters with the intraoperative anesthesia method to determine the postoperative recovery time. This makes the postoperative recovery time determination process more closely reflect the user's actual clinical scenario, thus comprehensively incorporating more key variables affecting postoperative recovery and effectively improving the accuracy of the determined postoperative recovery time.

[0033] The steps described above are explained in detail below.

[0034] First, regarding step 110, the meanings of the physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters involved in this application embodiment will be specifically explained.

[0035] Physiological characteristic parameters refer to quantitative or qualitative indicators used to reflect a user's current physical function, such as age, gender, and body mass index (BMI) (24 kg / m²). 2 ( ), cardiac function related indicators, respiratory function related indicators, other organ function indicators, anatomical and functional indicators of airway assessment, and functional status indicators of scale assessment.

[0036] These cardiac function-related indicators may include, but are not limited to: left ventricular ejection fraction (LVEF = 55%), transvalvular pressure gradient of the aortic valve (47 mmHg), and valve orifice area (AVA = 0.5 cm²). 21) Pulmonary artery pressure (systolic / diastolic / mean pressure mmHg), recent history of myocardial infarction, hypertension, atrial fibrillation, heart failure, New York Heart Association (NYHA) functional classification, and NT-proBNP levels. Respiratory function-related indicators may include, but are not limited to: respiratory function test results characterizing the type and severity of ventilation dysfunction, and blood gas analysis. Indicators of other organ functions may include, but are not limited to: renal function indicators, liver function indicators, and coagulation function indicators. Specifically, renal function indicators can be expressed by serum creatinine (88 μmol / L), a laboratory indicator reflecting visceral excretion function. Liver function indicators can be expressed by ALT and AST, laboratory indicators reflecting hepatocellular damage and liver function status. Coagulation function indicators can be expressed by quantitative indicators reflecting blood clotting ability, such as INR and platelets. Anatomical and functional indicators for airway assessment may include, but are not limited to: Mallampati classification, nail-mental distance, mouth opening, and neck mobility. The functional status indicators assessed by the scales may include, but are not limited to: cognitive function scores (Montreal Cognitive Assessment MoCA, Mini-Mental State Examination MMSE), Clinical Frailty Scale CFS, and Connecticut Collaborative Scale CCS.

[0037] Disease risk level parameters are indicators that assess the degree of risk to surgery and postoperative recovery based on a user's existing diseases, medical history, or comorbidities. These parameters can be obtained through diagnosis, medical history records, or standardized scoring. Specific parameters may include, but are not limited to: heart-related diseases and medical history, respiratory-related diseases, neurological diseases and medical history, other organ diseases, risk scores and comorbidity indicators, airway-related risk factors, and previous surgical history.

[0038] Cardiac-related diseases and histories may include, but are not limited to: history of myocardial infarction, hypertension, atrial fibrillation, heart failure, and New York Heart Association (NYHA) functional classification. Respiratory-related diseases may include, but are not limited to: chronic obstructive pulmonary disease (COPD) classification. Neurological diseases and histories may include, but are not limited to: stroke / TIA, carotid artery disease, cognitive impairment, dementia, and Parkinson's disease. Other organ diseases may include, but are not limited to: chronic renal insufficiency, Child-Pugh liver function, diabetes, and peripheral artery disease. Risk scores and comorbidity indicators may include the Charleson Comorbidity Index (CCIS score). Airway-related risk factors may include, but are not limited to: history of difficult airway and severity of OSA (obstructive sleep apnea). Past surgical history may include, but is not limited to: coronary artery bypass grafting (CABG), percutaneous coronary intervention (PCI), surgical aortic valve replacement (SAVR), and percutaneous transaortic valve replacement (TAVR).

[0039] Surgical parameters refer to indicators directly related to TAVR surgery, which may include surgical approach, valve type and size, special operational requirements, surgical risk assessment, expected surgical complexity, and anatomical features.

[0040] Surgical approaches may include, but are not limited to: transfemoral approach, transapical approach, and other approaches. Valve types may include, but are not limited to: balloon-expandable valves and self-expanding valves. Valve sizes may include, but are not limited to: 20mm, 23mm, 26mm, and 29mm. Special procedural requirements may include, but are not limited to: whether rapid ventricular pacing (RVP) is required, whether coronary artery protection or simultaneous PCI is required, and whether extracorporeal circulation (ECMO) is required. Surgical risk assessment can be determined by calculating the STS score. The expected surgical complexity and anatomical features can be classified according to the comprehensive CT anatomy risk of aortic valve disease, mainly used for preoperative assessment of patients with aortic valve disease to determine whether their anatomy is suitable for transcatheter aortic valve replacement, as detailed below.

[0041] Category A (Simple Lesion): The patient has a tricuspid valve, and all anatomical measurements are normal. This includes mild and uniform calcification located at the leaflet base; annular size within the appropriate range of 20-26 mm; coronary artery ostium height exceeding 12 mm; aortic sinus width greater than 30 mm; sinoaortic duct junction greater than 25 mm; ventricular diameter greater than 30 mm; annular angle less than 60 degrees; ascending aorta less than 40 mm; and good diameter of both iliac and femoral arteries. The anatomical structure of these patients is conducive to surgery, and the risk is relatively low.

[0042] Category B (Complex Lesions): Patients with bicuspid valves or anatomical structures outside the normal range. For example, severe calcification, annular enlargement or under-enlargement; coronary artery ostium height less than 12 mm with leaflets visible at the ostium level; annular angle greater than 60 degrees; ascending aortic dilation but less than 50 mm; mild stenosis or calcification at the access route, but with no more than three risk factors. These patients have complex anatomical structures, increasing the difficulty and risk of surgery.

[0043] Category C (High-Risk Complex Lesions): Patients with bicuspid valves or significant anatomical abnormalities, such as large or small valve annulus, have a high risk of coronary artery occlusion; severe leaflet calcification, severe transverse cardiac position, or severe dilatation of the ascending aorta is present; the access route has severe calcification, stenosis, and tortuosity, and there are more than 3 risk factors. These patients have highly complex anatomical structures and high surgical risks.

[0044] Category D (Contraindicated Lesions): Patients with extremely abnormal anatomy, such as extremely large or small valve annulus, extreme risk of coronary artery occlusion, extreme transverse heart position, extreme risk of paravalvular leak, extreme ascending aortic dilatation, or even no available access route. Due to anatomical issues, these patients are generally unsuitable for TAVR and related surgeries.

[0045] Anesthesia-related parameters refer to indicators related to anesthesia preparation, which may include, but are not limited to: anesthesia method, intraoperative medication, monitoring intensity, airway management method, respiratory management parameters, fluid management, intraoperative adverse events, anesthesia duration, and operation duration.

[0046] The anesthesia methods can include general anesthesia (GA) and non-general anesthesia. Non-general anesthesia methods can include monitored anesthesia management (MAC, i.e., local anesthesia combined with sedation (LA+S)) and local anesthesia alone (LA). Intraoperative medications can include, but are not limited to, the types and dosages of anesthetic drugs (sedatives, opioids, muscle relaxants) and vasoactive drugs (norepinephrine, epinephrine, vasopressin, etc.). Monitoring intensity can include, but is not limited to, standard monitoring; invasive arterial pressure (ABP) monitoring; central venous pressure (CVP) monitoring; pulmonary artery pressure (PAC) monitoring; continuous cardiac output monitoring; intraoperative transesophageal echocardiography (TEE) monitoring; depth of anesthesia (BIS) monitoring; or cerebral oxygen saturation (NIRS) monitoring. Airway management methods can include, but are not limited to, face mask ventilation, laryngeal mask airway (LMA) ventilation, and endotracheal intubation (ETT) ventilation. Respiratory management parameters may include, but are not limited to: ventilator parameters (tidal volume, respiratory rate, positive end-expiratory pressure, peak airway pressure, plateau pressure, inspiratory-to-expiratory ratio, inspiratory oxygen concentration, end-expiratory carbon dioxide partial pressure); pulse oxygen saturation (SpO2); blood gas analysis indicators, etc. Fluid management may include, but is not limited to: fluid intake and output, blood transfusion requirements, and transfusion-related information. Intraoperative adverse events may include, but are not limited to: the frequency and amplitude of severe hemodynamic fluctuations; occurrence of malignant arrhythmias; intolerance to right ventricular pressure (RVP); hypoxemia events; hypercapnia; frequency of emergency interventions (such as cardioversion, cardiac pacing, vasopressor injection, airway intervention, etc.); need for advanced support equipment (such as intra-aortic balloon pump (IABP), extracorporeal membrane oxygenation (ECMO), continuous renal replacement therapy (CRRT, etc.); need for emergency conversion to surgery; intraoperative new-onset atrioventricular block, paravalvular leak, vascular complications, etc. Anesthesia duration and surgical duration may include, but are not limited to: patient awakening time, endotracheal tube extubation time; postoperative analgesia regimen; and the patient's transfer to a general ward or intensive care unit (ICU) after surgery.

[0047] Medical team competence evaluation parameters refer to indicators used to reflect the experience and level of the medical team performing surgery. Specifically, they can be determined by at least one of the following indicators: surgeon's TAVR surgery experience, annual surgical volume of the medical institution, anesthesia fluctuation rate, collaboration score, and complex case ratio.

[0048] Among these, the surgeon's TAVR experience (number of TAVR procedures performed per year) is defined as ≥50 procedures per year for experienced surgeons (if the number exceeds 50, it is calculated as 50). The medical institution's annual surgical volume (total number of TAVR procedures per center per year) is defined as ≥100 procedures per year for experienced centers (if the number exceeds 100, it is calculated as 100). Anesthesia variability refers to the proportion of time during which the mean arterial pressure fluctuation exceeds 30% of the baseline value within the total anesthesia duration. Collaboration score refers to the proportion of cases covered by multidisciplinary team (MDT) collaboration (maximum score set at 100%). Complex case ratio refers to the proportion of complex cases such as bicuspid valve and high-risk coronary artery conditions within the total number of cases (when this proportion >20%, it reflects the team's technical comprehensiveness, and the portion exceeding 20% ​​is calculated as 0.2).

[0049] In some embodiments of this application, the medical team competence evaluation parameter Team_Score can be calculated using the following formula (1):

[0050]

[0051] It is worth noting that the evaluation parameters for medical team capabilities are out of 10 points, and teams with a score of 8 or higher are considered high-level teams.

[0052] In some embodiments of this application, the collected user parameters need to be processed as follows: continuous variables are standardized using the Z-score standardization method to eliminate dimensional differences; categorical variables are encoded using the one-hot encoding method to convert categorical information into a numerical form that can be used for calculation; and missing data are processed using an expert-driven missing value completion algorithm to ensure the integrity and usability of the data.

[0053] In some embodiments of this application, the expert experience-driven missing value completion algorithm can be implemented through the following steps 1 to 4.

[0054] Step 1: Construct conditional completion rules based on expert clinical experience. For example, when a patient is older than 75 years and left ventricular ejection fraction (LVEF) data is missing, the data is completed by referring to the average LVEF of patients in the same age group (e.g., 75-80 years old, or over 80 years old) to ensure that the data reflects the general physiological characteristics of this age group.

[0055] Step 2: Employ the similar case matching method. Construct a feature vector by extracting key patient features (such as underlying diseases, physiological indicators, surgical history, etc.), calculate the cosine similarity between the case to be completed and each case in the historical case database, and select the K most similar (e.g., K=5 or 10) matching cases. Then, perform a weighted average of the corresponding missing parameter values ​​of these K cases (weights can be allocated according to similarity, with higher similarity resulting in greater weights) as the completion result.

[0056] Step 3: Employ machine learning prediction. Using complete parameter data from historical cases as training samples, with known parameters as input features and the parameters to be completed as target variables, train a random forest regression model. For cases with missing parameters, input their known parameters into the trained model, and the model will predict and output estimated values ​​of the missing parameters.

[0057] Step 4: Integrate the completion results of the three methods mentioned above. Based on expert experience, assign weights to each method (e.g., rule-based completion emphasizes clinical consensus, similar case completion emphasizes individual matching, and machine learning emphasizes data patterns; these can be assigned weights of 40%, 30%, and 30% respectively). The final completion value is obtained through weighted calculation, balancing clinical rationality and data accuracy.

[0058] Secondly, regarding step 120, non-general anesthesia methods can specifically include monitored anesthesia management (i.e., local anesthesia combined with sedation) and simple local anesthesia. Simple local anesthesia refers to applying local anesthetic drugs to the corresponding localized site through techniques such as surface anesthesia, local infiltration anesthesia, regional block, or nerve block, temporarily and reversibly blocking the pain transmission of nerve fibers in that area, so that the patient remains fully conscious and can cooperate with commands while the pain at the surgical site disappears. Monitored anesthesia management (i.e., local anesthesia combined with sedation) can be somewhere between general and local anesthesia, using sedatives and analgesics to keep the patient in a mild to moderate sedative state while maintaining spontaneous breathing. General anesthesia is induced intravenously, such as with etomidate or cisatracurium, and maintained by inhalation, such as sevoflurane, to make the patient completely unconscious and able to control their breathing. For details, please refer to the introduction of anesthesia-related parameters in step 110 above, which will not be repeated here.

[0059] In some embodiments of this application, the intraoperative anesthesia method determination model includes an applicability assessment module, an applicability adjustment module, a complexity assessment module, and a decision module; based on this, such as Figure 2 As shown, step 120 above may specifically include steps 1201 and 1202.

[0060] Step 1201: Input the user parameters into the applicability assessment module, the applicability adjustment module, and the complexity assessment module respectively to obtain the first applicability score corresponding to the general anesthesia method and the second applicability score corresponding to the non-general anesthesia method output by the applicability assessment module, the applicability score adjustment factor output by the applicability adjustment module, and the complexity index output by the complexity assessment module. The first applicability score is used to characterize the applicability of the user using the general anesthesia method; the second applicability score is used to characterize the applicability of the user using the non-general anesthesia method; the applicability score adjustment factor is used to characterize the adjustment range of the applicability score corresponding to the general anesthesia method and the applicability score corresponding to the non-general anesthesia method; and the complexity index is used to characterize the complexity faced by the user in implementing the general anesthesia method and the non-general anesthesia method.

[0061] The first suitability score refers to the suitability score for general anesthesia, used to quantify the patient's suitability for general anesthesia. A higher first suitability score indicates a stronger suitability for general anesthesia. The second suitability score refers to the suitability score for non-general anesthesia, used to quantify the patient's suitability for non-general anesthesia. A higher second suitability score indicates a stronger suitability for non-general anesthesia. The suitability score adjustment factor refers to the correction parameter output by the suitability adjustment module, used to dynamically adjust the suitability scores for the above-mentioned general anesthesia methods and non-general anesthesia methods according to high-risk, preferred, or mandatory rules.

[0062] For example, user parameters are input into the applicability assessment module, and the applicability assessment module processes the user parameters to obtain a first applicability score corresponding to the general anesthesia method and a second applicability score corresponding to the non-general anesthesia method.

[0063] The applicability assessment module can calculate the applicability score GA_Score corresponding to the general anesthesia method using the following formula (2).

[0064] GA_Score=w1×Age_Factor+w2×Cardiac_Factor+w3×Respiratory_Factor+(2)w4×Cognitive_Factor+w5×Surgical_Factor+w6×Experience_Factor

[0065] Among them, Age_Factor is the age factor with a weight w1 = 0.2; Cardiac_Factor is the cardiac function factor with a weight w2 = 0.25; Respiratory_Factor is the respiratory function factor with a weight w3 = 0.2; Cognitive_Factor is the cognitive function factor with a weight w4 = 0.15; Surgical_Factor is the surgical factor with a weight w5 = 0.15; and Experience_Factor is the experience factor with a weight w6 = 0.05.

[0066] The older the age, the lower the applicability of general anesthesia. Based on this, the scoring rules for the age factor are as follows: 0.8 points for 70-79 years old; for every 10 years increase in age, the score decreases by 0.1 (e.g., 0.7 points for 80-89 years old, and 0.6 points for 90 years old and above).

[0067] The worse the cardiac function (the lower the left ventricular ejection fraction), the more general anesthesia is required. Based on this, the scoring rules for cardiac function factors are as follows: 0.8 points are assigned when the left ventricular ejection fraction (LVEF) is 40-49%; for every 10% decrease in LVEF, the score increases by 0.1 (e.g., 0.9 points for 30-39% and 1.0 point for <30%).

[0068] The worse the respiratory function, the lower the applicability of general anesthesia. Based on this, the scoring rules for the respiratory function factor are as follows: 0.8 points for normal respiratory function; 0.7 points for mild chronic obstructive pulmonary disease (COPD); 0.6 points for moderate COPD; and 0.5 points for severe COPD.

[0069] The more severe the cognitive impairment, the lower the applicability of general anesthesia. Based on this, the scoring rules for the cognitive function factor are as follows:

[0070] When using the Montreal Cognitive Assessment (MoCA): ≥26 points (normal cognition) are assigned 0.8 points, 18-25 points (mild impairment) are assigned 0.7 points, and <18 points (cognitive impairment) are assigned 0.6 points;

[0071] When using the Mini-Mental State Examination (MMSE, for those with secondary school education or above): ≥24 points (normal cognition) are assigned 0.8 points, 20-23 points (mild impairment) are assigned 0.7 points, 10-19 points (moderate impairment) are assigned 0.6 points, and <10 points (severe impairment) are assigned 0.5 points. (Note: For each increase in the severity of cognitive impairment, the score decreases by 0.1).

[0072] The more complex the surgery or the higher the anatomical risk, the more suitable general anesthesia is. Based on this, the scoring rules for the surgical factor are as follows: simple lesions are scored 0.5 points, complex lesions are scored 0.6 points, high-risk complex lesions are scored 0.7 points, and contraindicated lesions (requiring special treatment) are scored 0.8 points.

[0073] The more experienced the surgeon, the lower their dependence on general anesthesia. Based on this, the scoring rules for the experience factor are as follows: 0.8 points are awarded when the surgeon performs 50 surgeries per year; for every additional 25 surgeries per year, the score decreases by 0.1 (e.g., 0.7 points for 75 surgeries, and 0.6 points for 100 surgeries or more).

[0074] The suitability assessment module can calculate the suitability score LA_Score corresponding to non-general anesthesia methods using the following formula (3).

[0075] LA_Score=w1×Age_Factor+w2×Cooperation_Factor+w3×Frailty_Factor+(3)w4×Surgical_Factor+w5×Complication_Factor+w6×Recovery_Factor

[0076] Among them, Age_Factor is the age factor with a weight w1 = 0.15; Cooperation_Factor is the cooperation ability factor with a weight w2 = 0.2; Frailty_Factor is the frailty factor with a weight w3 = 0.15; Surgical_Factor is the surgical factor with a weight w4 = 0.25; Complication_Factor is the complication risk factor with a weight w5 = 0.15; and Recovery_Factor is the rapid recovery need factor with a weight w6 = 0.1.

[0077] The older the age, the more suitable non-general anesthesia becomes. Based on this, the scoring rules for the age factor are as follows: 0.8 points for 70-79 years old; and 0.1 points for every 10 years of age increase (e.g., 0.9 points for 80-89 years old, and 1.0 point for 90 years old and above).

[0078] The lower the patient's cooperation level, the lower the applicability of non-general anesthesia (the lower the score). Based on this, the scoring rules for the cooperation ability factor are as follows: according to the Connecticut Cooperation Score (CCS), 0.8 points are assigned to Level 1 (active cooperation), 0.7 points are assigned to Level 2 (passive cooperation), and 0.6 points are assigned to Level ≥3 (non-cooperation) (that is, the score decreases by 0.1 for each level increase).

[0079] The higher the degree of frailty, the greater the applicability of non-general anesthesia (increased score). Based on this, the scoring rules for the frailty factor are as follows: according to the Clinical Frailty Scale (CFS), <5 (normal state) is assigned 0.6 points, 5 (mild frailty) is assigned 0.7 points, 6 (moderate frailty) is assigned 0.8 points, 7 (severe frailty) is assigned 0.9 points, and ≥8 (extremely severe frailty, requiring careful evaluation of TAVR treatment benefit) is assigned 1.0 point (i.e., for each grade increase in CFS, the score increases by 0.1).

[0080] The more complex the surgery or the higher the anatomical risk, the lower the applicability of non-general anesthesia (the lower the score). Based on this, the scoring rules for the surgical factor are as follows: simple lesions (suitable for non-general anesthesia) are scored 0.8 points, complex lesions are scored 0.7 points, high-risk complex lesions are scored 0.6 points, and contraindicated lesions are scored 0.5 points.

[0081] The higher the risk of complications, the less suitable non-general anesthesia is (the lower the score). The scoring rules for complication risk factors are as follows: low-risk patients (STS score <4%) are assigned 0.8 points, intermediate-risk patients (STS score 4%–8%) are assigned 0.7 points, and high-risk patients (STS score >8%) are assigned 0.6 points.

[0082] The higher the demand for rapid recovery, the greater the applicability of non-general anesthesia (increased score). The scoring rules for the rapid recovery demand factor are as follows: high demand intensity (target postoperative hospital stay ≤ 3 days) is assigned 0.8 points, moderate intensity (target discharge in 4-7 days) is assigned 0.7 points, and low intensity (target hospital stay > 7 days) is assigned 0.6 points.

[0083] The following two examples will be used to illustrate the evaluation process of the two anesthesia methods.

[0084] The patient's user parameters are as follows: left ventricular ejection fraction 35%, moderate chronic obstructive pulmonary disease, normal cognitive function (MMSE score 24), good cooperation, mild weakness (HFRS score 3), surgical type is simple lesion, planned femoral artery approach, surgeon is experienced (150 surgeries per year), patient belongs to low-risk group (STS score 3%), and hopes to be discharged 3 days after surgery.

[0085] Calculation of general anesthesia suitability score:

[0086] Age_Factor = 0.7 (Score reduced for ages ≥ 80)

[0087] Cardiac_Factor = 0.9 (LVEF < 40%, increase score)

[0088] Respiratory Factor = 0.6 (Moderate COPD, lower the score)

[0089] Cognitive Factor = 0.8 (Normal Cognition)

[0090] Surgical Factor = 0.5 (Simple lesions via the femoral artery approach, not favoring general anesthesia)

[0091] Experience_Factor = 0.4 (The surgeon is highly experienced, so the score is lowered)

[0092] GA_Score=0.2×0.7+0.25×0.9+0.2×0.6+0.15×0.8+0.15×0.5+0.05×0.4=0.7

[0093] Calculation of the suitability score for non-general anesthesia:

[0094] Age_Factor = 0.9 (Elderly patients, suitable for non-general anesthesia)

[0095] Cooperation Factor = 0.8 (Good cooperation)

[0096] Frailty_Factor = 0.7 (Mildly Weak)

[0097] Surgical Factor = 0.9 (via femoral artery approach, suitable for non-general anesthesia)

[0098] Complication_Factor = 0.6 (STS 3%, relatively low risk)

[0099] Recovery_Factor = 0.8 (For faster recovery)

[0100] LA_Score=0.15×0.8+0.2×0.8+0.15×0.6+0.25×0.8+0.15×0.8+0.1×0.8=0.77

[0101] Based on the above experiments, it can be concluded that this patient is more suitable for non-general anesthesia.

[0102] After the user parameters are input into the suitability adjustment module, the module dynamically adjusts the suitability scores for general anesthesia and non-general anesthesia according to preset clinical rules, and finally outputs the suitability score adjustment factor. The specific rules are as follows:

[0103] High-risk general anesthesia rule: If the user parameters meet any of the following conditions, the applicability adjustment module will lower the applicability score of the general anesthesia method, that is, output an applicability score adjustment factor that lowers the applicability score of the general anesthesia method:

[0104] Patients aged >85 years with preoperative resting blood gas analysis (without oxygen) showing PaO2 <60 mmHg (suggesting respiratory failure);

[0105] Those with cognitive impairment and a significantly increased risk of delirium (preoperative MoCA score or MMSE score <10);

[0106] Patients with severe chronic obstructive pulmonary disease and a history of oxygen therapy dependence.

[0107] Preferred Non-General Anesthesia Rules: If the user parameters meet any of the following conditions, the applicability adjustment module will increase the applicability score of the non-general anesthesia method, that is, output an applicability score adjustment factor that increases the applicability score of the non-general anesthesia method:

[0108] The surgery was performed via the femoral artery approach and the patient cooperated well.

[0109] There is a special need for rapid recovery (such as the goal of being discharged on the same day, the next day, or within 48 hours after surgery);

[0110] The patient has a history of difficult intubation.

[0111] Forced General Anesthesia Rule: If the user parameters meet any of the following conditions, the applicability adjustment module will output an applicability score adjustment factor that forces the selection of general anesthesia, i.e., outputting an applicability score of 0 for non-general anesthesia methods is invalid, thus directly locking general anesthesia as the user's only option:

[0112] The surgery was performed via the transapical approach.

[0113] Based on a comprehensive assessment using CT anatomy, experts determined that the risk of switching anesthesia methods during surgery was extremely high.

[0114] The patient exhibited severe anxiety and was completely uncooperative (Connecticut Cooperation Scale CCS4).

[0115] Through the calculation of the above rules, the applicability score adjustment factor finally output by the applicability adjustment module will be used to correct the applicability score corresponding to the initial general anesthesia method and the applicability score corresponding to the non-general anesthesia method, so that the applicability score is more in line with the patient's individual risk and clinical needs.

[0116] User parameters are input into the complexity assessment module, which then calculates patient complexity, surgical complexity, and anesthesia technique complexity according to a preset multi-dimensional assessment system, and finally integrates and outputs a comprehensive complexity index. The specific process is as follows:

[0117] First, patient complexity is calculated. The complexity assessment module extracts age, comorbidities, cardiac function, and airway-related data from user parameters, quantifies and scores them according to the following rules, and then calculates the weighted average:

[0118] Age complexity: A score is assigned based on the patient's age range. For example, age <70 = 0 points, 70-79 = 0.7 points, 80-84 = 0.8 points, 85-89 = 0.9 points, and ≥90 = 1.0 points.

[0119] Comorbidity complexity: Based on the Charlson Comorbidity Index (CCI) classification and assignment, for example, CCI 0 = 0 points, CCI 1 = 0.2 points, CCI 2 = 0.4 points, CCI 3 = 0.6 points, CCI 4 = 0.8 points, CCI 5 and above = 1.0 points;

[0120] Cardiac function complexity: Assigned according to the left ventricular ejection fraction (LVEF) interval. For example, LVEF ≥ 50% (normal) = 0 points, LVEF 40-49% (mild decrease) = 0.5 points, LVEF 30-39% (moderate decrease) = 0.8 points, LVEF < 30% (severe decrease) = 1.0 point;

[0121] Airway complexity: Referring to the ARNE scoring system, a history of difficult intubation is assigned as follows: 0.4 points for a clearly documented history of difficult laryngoscopic exposure (Cormack-Lehane III / IV); 0.3 points for a Mallampati classification (Mallampati classification of only the soft palate base III or only the hard palate IV); 0.2 points for limited cervical spine mobility (limited flexion / extension <80° or limited rotation <70°); and 0.2 points for limited mandibular mobility (IID <5cm or the lower incisors not being able to move forward beyond the upper incisors, indicating limited mandibular protrusion). It is worth noting that if the total airway complexity score exceeds 1.0, it is counted as 1.0.

[0122] Ultimately, patient complexity = 0.3 × age complexity + 0.25 × comorbidity complexity + 0.25 × cardiac function complexity + 0.2 × airway complexity.

[0123] Secondly, the surgical complexity is calculated by first extracting relevant data on surgical approach, anatomical features, and operational requirements from user parameters through the complexity assessment module, and then quantifying and weighting the data according to the following rules:

[0124] Complexity of approach: The score is assigned based on the approach conditions assessed by CT anatomy. For example, a good iliofemoral artery diameter is 0 points; mild stenosis or calcification in the approach is 0.3 points; severe calcification, stenosis and tortuosity in the approach is 0.6 points; no available approach, transapical approach or other approach is 1.0 point.

[0125] Anatomical complexity: Based on the comprehensive risk classification of CT anatomy, if all factors meet the criteria for Category A, the score is 0. If all risk factors meet the criteria for Category B, the score is +0.1 points for each risk factor; if all risk factors meet the criteria for Category C, the score is +0.2 points for each risk factor; if all risk factors meet the criteria for Category D, the score is +0.3 points for each risk factor. If the total score exceeds 1.0 point, it is calculated as 1.0 point.

[0126] Operational complexity: Assigned values ​​based on the difficulty of the surgical procedure. For example, routine basic operation = 0 points, rapid ventricular pacing (RVP) required +0.3 points, coronary artery protection required +0.3 points, and extracorporeal circulation (ECMO) required +0.4 points.

[0127] Ultimately, the surgical complexity = 0.4 × approach complexity + 0.3 × anatomical complexity + 0.3 × operational complexity.

[0128] Furthermore, the complexity of anesthesia techniques is calculated by first extracting relevant data from user parameters through a complexity assessment module, then quantifying and scoring the data according to the following rules:

[0129] Monitoring complexity: Scoring is based on the monitoring method. For example, standard monitoring = 0 points, invasive arterial pressure (ABP) +0.2 points, central venous pressure (CVP) +0.2 points, floating catheter / pulmonary artery pressure (PAC) / continuous cardiac output monitoring +0.2 points, intraoperative transesophageal echocardiography (TEE) +0.2 points, depth of anesthesia (BIS) or cerebral oxygen saturation monitoring (NIRS) +0.2 points;

[0130] Management Complexity: The difficulty is assessed across three categories: anticipated hemodynamic management difficulty (basic anesthesia method + anticipated fluctuation risk + vasoactive drug requirement), respiratory management difficulty (airway management method + pulmonary function status), and anticoagulation / coagulation management difficulty (bleeding / thrombosis risk) (total score 1.0). Specifically, anticipated hemodynamic management difficulty (total score 0.5): Different anesthesia methods have different basic complexities: local anesthesia (LA) = 0 points, monitored anesthesia management (MAC) = 0.1 points, and general anesthesia (GA) = 0.2 points. Expected severe hemodynamic fluctuations / malignant arrhythmias / poor tolerance to rapid pacing / extreme sensitivity to volume status = +0.1 points (using random forest, predictor factors: severe heart failure, low EF, coronary artery disease, severe AS). Expected need for high-dose vasoactive drugs (types of vasoactive drugs): not needed = 0 points, 1 type = +0.1 points, 2 or more types = +0.2 points (using random forest, predictor factors: cardiac function, valvular disease, pulmonary hypertension). Respiratory management difficulty (total score 0.3): Different airway management methods have different basic complexities: face mask = 0 points, laryngeal mask airway (LMA =) 0.1 points, endotracheal intubation (ETT) = 0.2 points. Severe COPD / poor lung function / hypoxemia / hypercapnia = +0.1 points. Anticoagulation / coagulation management (total score 0.2): High bleeding risk (low platelet count, continued use of antiplatelet drugs, poor renal function) = +0.1 points. High thrombotic risk (atrial fibrillation, mechanical valves) = +0.1 points. Subsequently, management complexity = expected hemodynamic management difficulty + respiratory management difficulty + anticoagulation / coagulation management difficulty.

[0131] Then, the recovery complexity is calculated. The complexity assessment module accumulates scores based on factors such as postoperative analgesia requirements, extubation difficulty, complication risk, ICU transfer requirements, and transport risks. Specifically, the scores can be quantified and calculated according to the following rules:

[0132] Postoperative analgesia requirements: Simple intravenous analgesia = 0 points; Complex multimodal analgesia such as regional nerve block is required = +0.1 points.

[0133] Expected difficulty of extubation / recovery: Rapid extubation plan = 0 points; Expected delayed extubation = +0.1 points;

[0134] Intraoperative complication risk: High risk of paravalvular leakage = +0.1 points, High risk of atrioventricular block / pacemaker implantation = +0.1 points, High risk of vascular complications = +0.1 points.

[0135] Advanced monitoring (transfer to ICU) requirements: Transfer to general ward = 0 points; Planned transfer to ICU = +0.1 points; Unplanned transfer to ICU (due to complications) = +0.2 points.

[0136] Transport risks: hemodynamic instability and advanced support equipment, with IABP = +0.1 points, with ECMO = +0.1 points, with CRRT = 0.1 points.

[0137] Subsequently, the complexity of recovery equals the postoperative analgesia requirement, the expected difficulty of extubation / awakening, the risk of intraoperative complications, the need for advanced monitoring, and the risk of transport.

[0138] Recovery complexity: Scored based on factors such as postoperative analgesia needs, extubation difficulty, complication risk, ICU transfer needs, and transport risks (e.g., delayed extubation +0.1 points, transport with ECMO +0.1 points, etc.).

[0139] Ultimately, the complexity of anesthesia techniques = 0.4 × monitoring complexity + 0.3 × management complexity + 0.3 × recovery complexity.

[0140] The complexity assessment module integrates the above three complexity factors according to their weights to obtain a comprehensive complexity index. Specifically, the comprehensive complexity index = 0.4 × patient complexity + 0.35 × surgical complexity + 0.25 × anesthesia technique complexity, which serves as the final output.

[0141] Step 1202: Input the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index into the decision module to obtain the intraoperative anesthesia method output by the decision module.

[0142] The decision module is an algorithm model that integrates multi-dimensional parameters to make a final judgment. Specifically, it can be built based on a rule engine or machine learning algorithm. Its function is to comprehensively input the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index, and output the most suitable intraoperative anesthesia method for the patient.

[0143] For example, the first applicability score, second applicability score, applicability score adjustment factor, and comprehensive complexity index are converted into a numerical form recognizable by the decision module according to a preset format (e.g., standardized to the 0-1 range). The decision module first checks whether the applicability score adjustment factor contains mandatory rules, such as mandatory general anesthesia for transapical approach. If mandatory rules are included, the corresponding intraoperative anesthesia method is directly output without further calculation. If no mandatory rules exist, the decision module uses the applicability score adjustment factor to correct the first and second applicability scores. Subsequently, a weighted calculation is performed by combining the corrected applicability score and the complexity index. For example, a higher complexity index increases the weight of the general anesthesia method; then, the anesthesia method with the higher score is selected as the only recommended intraoperative anesthesia method.

[0144] Therefore, by integrating multi-dimensional quantitative indicators, dynamically corrected parameters, and risk complexity data, the limitations of a single scoring method are avoided, making the decision on anesthesia method more in line with the individual characteristics of the patient, and ultimately significantly improving the accuracy and reliability of intraoperative anesthesia method selection.

[0145] In some embodiments of this application, when it is detected that the applicability score adjustment factor does not contain a mandatory rule, such as Figure 3 As shown, step 1202 above may specifically include steps 12021 and 12022.

[0146] Step 12021: Input the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index into the decision module to obtain the first comprehensive score corresponding to the general anesthesia method and the second comprehensive score corresponding to the non-general anesthesia method; Step 12022: Through the decision module, determine the anesthesia method corresponding to the larger value between the first comprehensive score and the second comprehensive score as the intraoperative anesthesia method output by the decision module.

[0147] For example, the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index are input into the decision module. The decision module fuses the aforementioned parameters using a preset weighted summation algorithm or a machine learning model to calculate the first comprehensive score for general anesthesia and the second comprehensive score for non-general anesthesia. Subsequently, the decision module compares the first and second comprehensive scores and finally selects the anesthesia method with the highest score as the output result. For example, if the first comprehensive score is greater than the second comprehensive score, the general anesthesia method is recommended as the intraoperative anesthesia method and output; otherwise, the non-general anesthesia method is recommended as the intraoperative anesthesia method and output.

[0148] Therefore, by integrating the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index, the one-sidedness of a single indicator can be avoided, making the first comprehensive score and the second comprehensive score more in line with actual clinical needs, and significantly improving the accuracy and reliability of intraoperative anesthesia method selection.

[0149] Then, regarding step 130, the intraoperative anesthesia method plays a key role in the postoperative recovery process. By using the intraoperative anesthesia method as an important intermediate link in determining the user's intraoperative recovery time, it is possible to accurately predict the user's postoperative recovery time and provide decision support for clinical anesthesia management and postoperative rehabilitation.

[0150] In some embodiments of this application, such as Figure 4 As shown, step 130 above may specifically include steps 1301 to 1303.

[0151] Step 1301: Based on the intraoperative anesthesia method, select the target parameter type associated with the intraoperative anesthesia method from the parameter types. The parameter types include physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters.

[0152] Here, parameter type refers to the classification system of user parameters, including physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters. Target parameter type refers to the parameter type selected from the above parameter types that is related to the currently determined intraoperative anesthesia method, that is, the parameter classification that has a direct impact on postoperative recovery under this anesthesia method.

[0153] For example, based on the association between preset anesthesia methods and parameter types, the reference type associated with the intraoperative anesthesia method is determined as the target parameter type.

[0154] Step 1302: Based on the intraoperative anesthesia method, select target parameters associated with the intraoperative anesthesia method from the target parameter set corresponding to the target parameter type.

[0155] The target parameter set refers to the specific parameters included in each target parameter type. For details, please refer to the explanation of user parameters in step 110 above; no specific limitations are provided here. Target parameters are specific parameters that directly affect the postoperative recovery time of the current anesthesia method, further selected from the target parameter set based on the target parameter type.

[0156] For example, the parameter associated with the parameter type is determined as the target parameter based on the parameter type and the parameter association.

[0157] Step 1303: Input the target parameters into the postoperative recovery time determination model to obtain the postoperative recovery time output by the postoperative recovery time determination model.

[0158] Among them, the postoperative recovery time determination model refers to a prediction model based on machine learning. By training the correlation between target parameters and postoperative recovery time in historical case data, it can achieve quantitative prediction of the postoperative recovery time of individual patients.

[0159] Therefore, by using a two-level screening of parameter types and specific parameters, the parameters input into the postoperative recovery time determination model can focus on the key factors most relevant to the currently determined intraoperative anesthesia method, avoiding interference from irrelevant parameters, reducing the computational load of the postoperative recovery time determination model, and improving the relevance of the input data. This differentiated postoperative recovery time determination model that combines intraoperative anesthesia methods makes the prediction of postoperative recovery time more in line with the characteristics of different anesthesia methods, and significantly improves the accuracy of the postoperative recovery time output by the intraoperative recovery time determination model.

[0160] In some embodiments of this application, the postoperative recovery time determination model includes a first postoperative recovery time determination model corresponding to non-general anesthesia methods and a second postoperative recovery time determination model corresponding to general anesthesia methods; based on this, such as Figure 5 As shown, step 1303 can specifically include steps 13031 and 13032.

[0161] Step 13031: When the intraoperative anesthesia method is not general anesthesia, input the target parameter into the first postoperative recovery time determination model to obtain the postoperative recovery time output by the first postoperative recovery time determination model.

[0162] The first postoperative recovery time determination model includes a first immediate recovery time determination module, a first short-term recovery time determination module, and a first long-term recovery time determination module. Immediate recovery prediction targets recovery from 0-2 hours post-surgery, including awakening / consciousness recovery time and extubation time. Short-term recovery prediction targets recovery from 2-24 hours post-surgery, including cognitive function recovery time, pain control target achievement time, and basic living activities recovery time. Long-term recovery prediction targets recovery from 1-7 days post-surgery, including full mobility recovery time and discharge preparation time. Therefore, step 13031 may specifically include: inputting the target parameters into the first immediate recovery time determination module, the first short-term recovery time determination module, and the first long-term recovery time determination module respectively, and obtaining the output of the first immediate recovery time determination module. The first immediate recovery time, the first short-term recovery time, and the first long-term recovery time are determined by the first immediate recovery time and the first short-term recovery time, respectively. The first immediate recovery time is used to characterize the time required for a user to complete a first type of immediate recovery event within a first preset time window after surgery when using non-general anesthesia during the operation. The first short-term recovery time is used to characterize the time required for a user to complete a first type of short-term recovery event within a second preset time window after surgery when using non-general anesthesia during the operation. The first long-term recovery time is used to characterize the time required for a user to complete a first type of long-term recovery event within a third preset time window after surgery when using non-general anesthesia during the operation. The first immediate recovery time, the first short-term recovery time, and the first long-term recovery time are determined as the postoperative recovery time output by the first postoperative recovery time determination model.

[0163] In the embodiments of this application, the target parameters associated with non-general anesthesia methods may include, but are not limited to: type and dosage of sedative drugs, patient compliance, level of weakness, intraoperative monitoring parameters, pain sensitivity, analgesia regimen, psychological status score, postoperative nursing quality evaluation parameters, local anesthesia effect, time to achieve pain control target, healing grade of vascular puncture site, early mobilization intervention plan, baseline weakness, Hb (hemoglobin) change trend, NT-proBNP (cardiac function marker) level, perioperative vascular complications, and level of social support. The first pre-set postoperative time window refers to 0-2 hours postoperatively, with the first type of immediate recovery event being the recovery of consciousness. The second pre-set postoperative time window refers to 2-24 hours postoperatively, with the first type of short-term recovery event being the achievement of pain control target and the recovery of basic living abilities. The third pre-set postoperative time window is 1-7 days postoperatively, with the first type of long-term recovery event being the full recovery of mobility and discharge preparation.

[0164] The first immediate recovery time determination module, namely the consciousness recovery time determination module, can use a random forest regression analyzer. The type and dosage of sedative drugs, the patient's level of cooperation, the level of weakness, and intraoperative monitoring parameters are input into the random forest regression analyzer. The random forest regression analyzer processes the type and dosage of sedative drugs, the patient's level of cooperation, the level of weakness, and the intraoperative monitoring parameters, and outputs the consciousness recovery time within 0-2 hours after the operation.

[0165] The first short-term recovery time determination module can specifically include a pain control achievement time determination component and a basic activities of living (ADL) recovery time determination component. The pain control achievement time determination component can use an XGBoost regression analyzer. Pain sensitivity, analgesia regimen, psychological status score, and postoperative care quality evaluation parameters are input into the XGBoost regression analyzer. The XGBoost regression analyzer processes these parameters and outputs the pain control achievement time within 2-24 hours post-surgery. The basic activities of living (ADL) recovery time determination component can use a gradient boosting regression analyzer. Local anesthesia effect, pain control achievement time, vascular puncture site healing level, and early activity intervention plan are input into the gradient boosting regression analyzer. The gradient boosting regression analyzer processes these parameters and outputs the basic activities of living (ADL) recovery time within 2-24 hours post-surgery.

[0166] The first long-term recovery time determination module can specifically include a component for determining the full mobility recovery time and a component for determining the accurate discharge time. The full mobility recovery time determination component can use a gradient boosting regression analyzer. The analyzer processes the local anesthesia effect, time to achieve pain control targets, vascular puncture site healing level, and early mobilization intervention plan, outputting the full mobility recovery time within 1-7 days post-surgery. The accurate discharge time determination component can use LightGBM. The analyzer processes the baseline weakness, Hb (hemoglobin) trend, NT-proBNP (cardiac function marker) level, perioperative vascular complications, and social support level, outputting the accurate discharge time within 1-7 days post-surgery.

[0167] For example, for users who undergo non-general anesthesia, the aforementioned first postoperative recovery time determination model is used to process the user's target parameters, outputting a combination of the first immediate recovery time, the first short-term recovery time, and the first long-term recovery time. For example, immediate recovery prediction (0-2 hours): recovery time of consciousness: predicted as 5 minutes (confidence interval: 3.5-6 minutes); short-term recovery prediction (2-24 hours): time to achieve pain control target: predicted as 4 hours (confidence interval: 3-6 hours); recovery time of basic activities of daily living: predicted as 6 hours (confidence interval: 4-10 hours); long-term recovery prediction (1-7 days): time to full mobility recovery: predicted as 2 days (confidence interval: 1.5-3 days); discharge preparation time: predicted as 2.5 days (confidence interval: 2-3.5 days).

[0168] Step 13032: If the intraoperative anesthesia method is general anesthesia, input the target parameter into the second postoperative recovery time determination model to obtain the postoperative recovery time output by the second postoperative recovery time determination model.

[0169] The second postoperative recovery time determination model includes a second immediate recovery time determination module, a second short-term recovery time determination module, and a second long-term recovery time determination module. Based on this, step 13032 can specifically include: inputting the target parameters into the second immediate recovery time determination module, the second short-term recovery time determination module, and the second long-term recovery time determination module respectively, to obtain the second immediate recovery time output by the second immediate recovery time determination module, the second short-term recovery time output by the second short-term recovery time determination module, and the second long-term recovery time output by the second long-term recovery time determination module. The second immediate recovery time characterizes the time required for the user to complete a second type of immediate recovery event within a second preset time window after surgery when using general anesthesia during surgery. The second short-term recovery time characterizes the time required for the user to complete a second type of short-term recovery event within a second preset time window after surgery when using general anesthesia during surgery. The second long-term recovery time characterizes the time required for the user to complete a second type of long-term recovery event within a third preset time window after surgery when using general anesthesia during surgery. The second immediate recovery time, the second short-term recovery time, and the second long-term recovery time are determined as the postoperative recovery time output by the second postoperative recovery time determination model.

[0170] In the embodiments of this application, the target parameters associated with the general anesthesia method may include, but are not limited to: age, type and dosage of anesthetic drugs, operation time, intraoperative hemodynamic parameters, baseline respiratory function (vital capacity, blood gas analysis), intraoperative respiratory management parameters, dosage and duration of neuromuscular blocking drugs, recovery quality score, baseline cognitive function (preoperative MMSE score), anesthesia depth monitoring data (BIS value), intraoperative cerebral oxygen saturation, inflammatory marker levels, preoperative activity ability score, pain management effect score, hemodynamic stability score, early rehabilitation intervention, baseline cardiac function (EF value), preoperative serum albumin level, STS score, degree of mitral regurgitation, surgical approach, new postoperative conduction block, walking distance within 24 hours postoperatively, duration of mechanical ventilation, ICU stay, perioperative complications (infection, renal insufficiency, stroke), and level of social support. The first pre-set time window after surgery refers to 0-2 hours after surgery. The second category of immediate recovery events includes awakening events and extubation events. The second pre-set time window after surgery refers to 2-24 hours after surgery. The second category of short-term recovery events includes cognitive recovery events and basic living activities recovery events. The third pre-set time window after surgery is 1-7 days after surgery. The second category of long-term recovery events includes full recovery of mobility events and discharge preparation events.

[0171] For example, the second immediate recovery time determination module may specifically include an awakening time determination component and an extubation time determination component. The awakening time determination component may employ a random forest regression analyzer. Age, type and dosage of anesthetic drugs, surgical time, and intraoperative hemodynamic parameters are input into the random forest regression analyzer. The random forest regression analyzer processes the age, type and dosage of anesthetic drugs, surgical time, and intraoperative hemodynamic parameters, outputting the awakening time within 0-2 hours postoperatively. The extubation time determination component may employ a gradient boosting regression analyzer. Baseline respiratory function, intraoperative respiratory management parameters, neuromuscular blocking drug dosage and duration, and awakening quality score are input into the gradient boosting regression analyzer. The gradient boosting regression analyzer processes the baseline respiratory function, intraoperative respiratory management parameters, neuromuscular blocking drug dosage and duration, and awakening quality score, outputting the extubation time within 0-2 hours postoperatively.

[0172] The second short-term recovery time determination module can specifically include a cognitive recovery time determination component and a basic activities of living (ADL) recovery time determination component. The cognitive recovery time determination component uses an XGBoost regression analyzer. Baseline cognitive function, anesthesia depth monitoring data, intraoperative cerebral oxygen saturation, and inflammatory marker levels are input into the XGBoost regression analyzer. The XGBoost regression analyzer processes these data and outputs the cognitive function recovery time within 2-24 hours postoperatively. The ADL recovery time determination component uses a LightGBM regression analyzer. Preoperative activity ability scores, pain management effectiveness scores, hemodynamic stability scores, and early rehabilitation intervention information are input into the LightGBM regression analyzer. The LightGBM regression analyzes these data and outputs the ADL recovery time within 2-24 hours postoperatively.

[0173] The second long-term recovery time determination module can specifically include a component for determining the full mobility recovery time and a component for determining the accurate discharge time. The full mobility recovery time determination component can use a LightGBM regression analyzer. Preoperative serum albumin levels, STS scores, degree of mitral regurgitation, surgical approach, new postoperative conduction block, and walking distance within 24 hours postoperatively are input into the LightGBM regression analyzer. The LightGBM regression analyzer processes these factors and outputs the full mobility recovery time within 1-7 days postoperatively. The accurate discharge time determination component can use a random forest model. Mechanical ventilation duration, ICU stay time, perioperative complications, and social support levels are input into the random forest model. The random forest model processes these factors and outputs the accurate discharge time within 1-7 days postoperatively.

[0174] For example, for users undergoing general anesthesia, the aforementioned second postoperative recovery time determination model is used to process the user's target parameters, outputting a combination of the second immediate recovery time, the second short-term recovery time, and the second long-term recovery time. For example, immediate recovery prediction (0-2 hours): recovery time of consciousness: predicted as 5 minutes (confidence interval: 3.5-6 minutes); short-term recovery prediction (2-24 hours): time to achieve pain control target: predicted as 4 hours (confidence interval: 3-6 hours); recovery time of basic activities of daily living: predicted as 6 hours (confidence interval: 4-10 hours); long-term recovery prediction (1-7 days): time to full mobility recovery: predicted as 2 days (confidence interval: 1.5-3 days); discharge preparation time: predicted as 2.5 days (confidence interval: 2-3.5 days).

[0175] Therefore, by screening target parameters associated with intraoperative anesthesia methods based on the differentiated characteristics of non-general anesthesia and general anesthesia methods, interference from irrelevant parameters can be avoided, improving the processing efficiency of the first recovery time determination model and the second recovery time determination model. Furthermore, this phased and anesthesia-type-specific model design achieves full-cycle coverage from immediate postoperative recovery to long-term recovery, which can enhance the clinical reference value of postoperative recovery time.

[0176] It is worth noting that, in order to convert the intraoperative anesthesia method into a digital signal that the postoperative recovery time determination model can recognize, a multi-dimensional feature encoding can be performed on the intraoperative anesthesia method. Each number corresponds to an anesthesia operation feature. For example, a three-digit code can be used to represent the intraoperative anesthesia method: the first digit indicates whether general anesthesia was used (0 = no, 1 = used), the second digit indicates whether sedation drugs were used (0 = no, 1 = used), and the third digit indicates whether local anesthesia was used (0 = no, 1 = used). Based on this, the code for monitored anesthesia management (MAC: LA+S) is [0, 1, 0]; the code for local anesthesia alone (LA) is [0, 0, 1]; and the code for general anesthesia (GA) is [1, 0, 0].

[0177] In some embodiments of this application, an end-to-end joint training method can be used to simultaneously optimize the intraoperative anesthesia method determination model and the postoperative recovery time determination model. Specifically, a joint loss function can be defined based on intraoperative anesthesia method loss, postoperative recovery time loss, consistency loss, and expert knowledge loss. For example, the joint loss function can be represented by the following formula (4).

[0178] Total_Loss=0.3×Anesthesia_Loss+0.4×Recovery_Loss (4)

[0179] +0.2×Consistency_Loss+0.1×Expert_Loss

[0180] Among them, Anesthesia_Loss is the classification cross-entropy loss of intraoperative anesthesia mode prediction, Recovery_Loss is the mean squared error loss of recovery time prediction, Consistency_Loss is the loss to ensure consistency between intraoperative anesthesia mode prediction and postoperative recovery time prediction, and Expert_Loss is the loss of incorporating expert knowledge.

[0181] Anesthesia_Loss can use classification cross-entropy loss to quantify the intraoperative anesthesia method and determine the degree of deviation between the model-recommended intraoperative anesthesia method and historical high-quality clinical decisions, such as the intraoperative anesthesia method recommended by senior physicians. The greater the deviation, the higher the loss value. Recovery_Loss can use mean squared error loss to calculate the postoperative recovery time and determine the squared difference between the model-predicted postoperative recovery time and the patient's actual recovery time. The greater the time difference, the higher the loss value. Consistency_Loss is used to ensure that the intraoperative anesthesia method and the postoperative recovery time are logically consistent. Specifically, the consistency loss can be expressed by the following formula (5).

[0182] Consistency_Loss=α×|Predicted_Anesthesia_Effect- (5)Actual_Recovery_Pattern|+β×|Expert_Expectation-Model_Prediction|

[0183] Wherein, Predicted_Anesthesia_Effect is the predicted recovery effect based on the intraoperative anesthesia method, Actual_Recovery_Pattern is the actual observed recovery pattern, Expert_Expectation is the expert's expected postoperative recovery time based on the intraoperative anesthesia method, Model_Prediction is the recovery time predicted by the postoperative recovery time determination model, and α and β are balance coefficients, with values ​​of 0.6 and 0.4, respectively.

[0184] Expert_Loss transforms medical consensus into mathematical constraints, preventing risky decisions at the source, such as the use of contraindicated drugs or dosage over-dosing, thus mitigating the risk of purely data-driven models violating common medical sense. Specifically, expert knowledge loss applies loss values ​​to decisions that do not conform to clinical norms through pre-defined rules, ensuring that the model's behavior is consistent with medical principles and practical consensus. This can include, but is not limited to, the following dimensions:

[0185] 1. Drug dosage safety boundary constraints are used to ensure that the dosages of anesthetic drugs and vasoactive drugs recommended by the model are within the physiologically safe and acceptable range, and to avoid the model from outputting dangerous anesthetic drug dosages that may cause circulatory inhibition, hypotension or even cardiac arrest due to training data bias or noise suggestions. At the same time, it prevents the recommendation of dangerous vasoactive drug dosages that may cause circulatory hyperdynamic state and microcirculatory failure.

[0186] 2. Drug contraindications or relative contraindications are used to penalize drug use scenarios with known risks of serious adverse reactions, preventing the model from recommending treatments that may trigger severe allergic reactions, acute myocardial ischemia, or malignant arrhythmias. For example, for patients with known severe aortic stenosis and coronary artery disease, if the model recommends the use of adrenaline, which may induce tachycardia, a high penalty will be triggered to prevent such decisions that violate contraindications.

[0187] 3. Physiological parameter adaptation constraints ensure that anesthesia decisions (such as drug selection and ventilation parameter settings) match the patient's current user parameters, guaranteeing that intraoperative decision-making is feasible and safe under the patient's current pathophysiological environment. For example, when a patient has severe hypovolemic shock, if the model prioritizes vasodilatory anesthetics (such as propofol) instead of prioritizing volume resuscitation, it will be judged as not conforming to the physiological adaptation principle, resulting in a loss penalty.

[0188] 4. Logical consistency constraints between decision-making and prediction results are used to ensure that the mapping relationship of "intraoperative anesthesia method → ​​postoperative recovery time" learned by the model conforms to known pharmacological and physiological principles, thereby improving the interpretability and reliability of the model. For example, if the model selects general anesthesia but its predicted recovery time is shorter than the expected recovery time of similar patients without general anesthesia (violating the basic principle that general anesthetic drugs have a longer metabolic cycle), a logical consistency loss penalty will be triggered to calibrate the correlation logic between decision-making and prediction within the model.

[0189] In some embodiments of this application, the above-described end-to-end joint training method can be implemented using a multi-objective optimization strategy, with the specific steps as follows:

[0190] Training data preparation: Collect historical case data, including: Input features: User parameters, including physiological characteristic parameters to characterize the user's physiological state, disease risk level parameters to characterize the user's disease, surgery-related parameters, anesthesia-related parameters, and medical team competence evaluation parameters; Label data: Anesthesia method selected by senior physicians for calculating Anesthesia_Loss, actual patient recovery time for calculating Recovery_Loss, and the expected relationship between expert's intraoperative anesthesia method and postoperative recovery time for calculating Consistency_Loss and Expert_Loss.

[0191] Loss function calculation and backpropagation: After each output of the intraoperative anesthesia method and postoperative recovery time in the above combined model, the combined loss is calculated and the parameters are updated according to the above formulas (4) and (5).

[0192] Multi-objective optimization and balancing: Using the Pareto optimal solution set optimization method, the following four objectives are balanced:

[0193] Accuracy objective (weight 0.4): Minimize the weighted sum of Anesthesia_Loss and Recovery_Loss in the above formula (4) to ensure that the anesthesia method and recovery time output by the model are consistent with the actual clinical results.

[0194] Interpretability objective (weight 0.2): Maximize the clarity of the decision path and the uniformity of the feature importance distribution, that is, maximize the Expert_Loss in the above formula (4) so ​​that the model decision logic can be understood by clinicians and avoid black box prediction; Specifically, the clarity of the decision path can be made traceable by constraining the model structure (such as using an interpretable tree model) to make the path of "why choose [0, 1, 0] (MAC) instead of [0, 0, 1] (LA)" traceable; while the uniformity of feature importance can be achieved by adding a regularization term to the loss to avoid the model from over-relying on a single feature, such as choosing LA just because the patient is young, ignoring clinically relevant features such as operation time and underlying diseases, and the standard for a reasonable feature distribution can be defined by anesthesiologists through experience.

[0195] Clinical utility objective (weight 0.3): Maximize the combination of decision support value and risk stratification value, that is, maximize Consistency_Loss in the above formula (4) to ensure that the model output can directly guide clinical operation and distinguish the patient risk level; Specifically, decision support value can ensure that the anesthesia method and recovery time output by the model are consistent with clinical logic through consistency loss. For example, if the model predicts [0, 1, 0] (MAC), the corresponding recovery time needs to be slightly longer than [0, 0, 1] (LA); risk stratification value can be achieved by consistency loss, allowing the model to group patients according to the risk combination of intraoperative anesthesia method and postoperative recovery time. For example, selecting LA and a long recovery time of less than 2 days is a low-risk group, and selecting MAC and a long recovery time of more than 5 days is a high-risk group. The stratification results can be directly used for clinical resource allocation.

[0196] Robustness objective (weight 0.1): Cross-validation ensures model stability across different datasets while enhancing resistance to noisy data. By maximizing stability and robustness, the model maintains stable output across different hospital datasets and noisy data. Specifically, cross-validation stability can be improved by adjusting hyperparameters (such as the model learning rate) through K-fold cross-validation to avoid the model only adapting to conscious and sedated data from a particular hospital. For example, if a hospital has a high MAC (consciousness-to-sedation) ratio, the model might fail in hospitals with low MAC ratios. Early stopping strategies are also used to prevent overfitting. Noise robustness can be improved by joint training to share features, allowing the model to learn the essential characteristics of conscious and sedated data and reduce the impact of noisy data. In addition, data augmentation can be performed by actively adding noise, or L1 / L2 regularization can be used to penalize model complexity, preventing the model from becoming overly sensitive to noisy data.

[0197] By combining Pareto optimization strategies, a balance is found between accuracy, interpretability, clinical applicability, and robustness, making the models for determining intraoperative anesthesia methods and postoperative recovery time both accurate and easy to understand clinically, while also adapting to the data characteristics of different hospitals.

[0198] During training, the weights of each loss term are dynamically adjusted to minimize the joint loss function.

[0199] Joint training avoids the shortcomings of the models for determining intraoperative anesthesia methods and postoperative recovery time operating independently. Consistency_Loss ensures logical consistency between the two models, thereby improving overall clinical applicability.

[0200] In some embodiments of this application, during the use of the intraoperative anesthesia method determination model and the postoperative recovery time determination model, actual clinical surgical results can be collected, such as the final anesthesia method used, intraoperative drug adjustment plan, and postoperative recovery time data, such as the actual duration of each recovery stage. These data are then compared and analyzed in a structured manner with the model's prediction results to quantify prediction biases, such as the accuracy rate of intraoperative anesthesia method recommendations and the prediction error of postoperative recovery time. A model performance baseline and monitoring index system are established to track key parameters such as prediction accuracy and error fluctuation range in real time. When core indicators are detected, such as the intraoperative anesthesia method prediction accuracy decreasing beyond a preset threshold, or the postoperative recovery time prediction error exceeding the acceptable range for multiple consecutive batches, the update process for the intraoperative anesthesia method determination model and the postoperative recovery time determination model is automatically triggered. The update of the intraoperative anesthesia method determination model and the postoperative recovery time determination model adopts an incremental learning algorithm, periodically incorporating new case data to update the models while retaining historical training knowledge. Multi-center validation ensures the consistency of the intraoperative anesthesia method determination model and the postoperative recovery time determination model under different medical scenarios, and expert feedback is used to correct the Expert_Loss constraint rules. In this way, through incremental learning and multi-center validation, the model can continuously absorb new clinical experience, improve prediction accuracy and safety margins, and extend its lifespan.

[0201] In some embodiments of this application, such as Figure 6 As shown, the above method for determining postoperative recovery time may further include steps 210 to 230.

[0202] Step 210: Input the intraoperative anesthesia method and user parameters into the complication risk assessment model to obtain the complication type output by the complication risk assessment model and the risk probability of the user having the corresponding complication type.

[0203] The complication risk assessment model can identify high-risk complication types, such as respiratory failure after general anesthesia and vascular complications after local anesthesia, using classification algorithms such as logistic regression and random forest. It can also calculate the probability of each complication, such as an 8% risk of respiratory failure in patients under general anesthesia. This allows for the precise identification of potential complications that may affect postoperative recovery time.

[0204] For example, the construction of the complication risk assessment model fully relies on guidelines, literature, and expert experience. First, it selects model input parameters that are highly correlated with the prediction results from the user parameters. These input parameters can specifically include patient baseline dimensions (e.g., age, gender, underlying diseases), laboratory test and CT imaging data dimensions (e.g., liver and kidney function indicators, lung imaging features), hospital level, and surgeon experience. This complication risk assessment model uses classification algorithms such as logistic regression and random forest to not only identify high-risk complication types, such as respiratory failure after general anesthesia and vascular complications after local anesthesia, but also to accurately calculate various anesthesia-related risks, including postoperative cognitive impairment risk, postoperative nausea and vomiting risk, respiratory complication risk, and cardiovascular complication risk. For different complications, the complication risk assessment model will perform risk assessment based on their key characteristics. For example, when assessing the risk of postoperative cognitive impairment (POCD), input features may include the preoperative MMSE difference, duration of intraoperative BIS <40, hypotension, and duration of NIRS-monitored cerebral oxygen saturation <50%; when assessing the risk of postoperative nausea and vomiting (PONV), input features may include reference Apfel scores (female / non-smoker / history of PONV / postoperative opioid use), and dosage of 5-HT3 receptor antagonists; when assessing the risk of respiratory complications, input features may include preoperative SpO2, predicted FEV1%, and intraoperative respiratory parameters; when assessing the risk of cardiovascular complications, input features may include the number of intraoperative ST segment deviations >1 mm and peak vasoactive drug score (VIS). Using these key input features, the complication risk assessment model calculates the risk probability and confidence interval for each complication.

[0205] Step 220: Input the complication type, risk probability, and postoperative recovery time into the anesthesia effect determination model to obtain the indirect effect value output by the anesthesia effect determination model; the indirect effect value is used to characterize the degree of influence of intraoperative anesthesia mode on postoperative recovery time through complications.

[0206] The indirect effect value represents the additional extension of postoperative recovery time caused by a specific anesthetic method leading to a particular complication. The anesthetic impact determination model can calculate the indirect effect value using a regression algorithm based on the complication type output in step 210 and the risk probability of the user experiencing that complication type, combined with the correlation between complications and recovery time in clinical data. For example, if a patient under general anesthesia experiences respiratory failure (risk probability 10%), the indirect effect value could be a 10% extension of recovery time. In this way, the potential impact of complications is transformed into a calculable quantitative value, establishing an indirect correlation path between intraoperative anesthetic methods and recovery time.

[0207] Step 230: Adjust the postoperative recovery time using the indirect effect value to obtain the target recovery time, which is the postoperative recovery time of the user under the influence of complications.

[0208] For example, the target recovery time can be determined based on the initial postoperative recovery time and the indirect effect value of each complication. For each complication output in step 210 (such as respiratory failure, acute kidney injury, etc.), the corresponding indirect effect value is calculated by combining its risk probability and the proportion of the complication's impact on recovery time in clinical data. For example, the indirect effect value is the product of the risk probability of the proportion of prolonged recovery time caused by the complication.

[0209] Respiratory failure: The risk probability is 10%, and the baseline proportion of prolonged recovery time is 100% (i.e., the prolongation is the same as the initial time). Therefore, the indirect effect value is 100% × 10% = 10%.

[0210] Acute kidney injury: Risk probability 5%, baseline prolongation rate 80%, then indirect effect value = 80% × 5% = 4%.

[0211] Subsequently, the indirect effect values ​​of all complications are summed up and multiplied by the initial postoperative recovery time to obtain the final target recovery time. Using the initial recovery time = 5 days, the indirect effect value of respiratory failure = 10%, and the indirect effect value of acute kidney injury = 4%, then: target recovery time = 5 × (1 + 10% + 4%) = 5 × 1.14 = 5.7 days.

[0212] Therefore, by using indirect effect values ​​to convert the impact of intraoperative anesthesia methods on complications, and the impact of complications on postoperative recovery time, into specific numerical values, medical staff can intuitively understand the impact of complications on recovery time. The revised target recovery time integrates direct recovery trends and complication risks, improving the clinical applicability of the prediction and providing refined guidance for anesthesia decisions and complication management.

[0213] In some embodiments of this application, such as Figure 7 As shown, the above method for determining postoperative recovery time may further include steps 310 to 350.

[0214] Step 310: Based on the user's risk probability of developing complications, determine the target risk probability interval that matches the risk probability from multiple preset risk probability intervals corresponding to the complications.

[0215] The risk probability range can be determined based on the probability range preset by clinical data, such as low risk: 0-5%, medium risk: 6-20%, high risk: 21% or more. Different ranges can be set for different complications. For example, the high risk range for respiratory failure after general anesthesia can be set to ≥15%, while the high risk range for vascular complications after local anesthesia can be set to ≥10%.

[0216] For example, the user's risk probability of developing complications is matched with multiple preset risk probability intervals. For instance, for respiratory failure complications, 12% falls within the medium risk (6-20%) interval, and this risk probability interval is the target risk probability interval.

[0217] Step 320: Determine the risk level corresponding to the target risk probability interval as the user's first risk level. The first risk level is used to characterize the user's risk level of developing complications.

[0218] The first risk level refers to a qualitative label defined based on the target risk probability range, such as low risk, medium risk, and high risk, which directly reflects the likelihood of complications occurring.

[0219] For example, based on the preset mapping relationship between risk probability ranges and risk levels, the risk level mapped to the target risk probability range is determined as the first risk level. For instance, if the target risk probability range is 6-20%, then the first risk level is medium risk.

[0220] Step 330: Based on the target recovery time, determine the target recovery time interval that matches the target recovery time from multiple preset recovery time intervals.

[0221] The recovery time range refers to the preset time range based on the type of surgery and the method of anesthesia. For example, the recovery time of full mobility after general anesthesia for TAVR patients is: low risk ≤ 1.8 days, 1.8 days < medium risk ≤ 2 days, and high risk > 2 days.

[0222] For example, the target recovery time is matched with multiple preset recovery time intervals, and the recovery time interval mapped by the target recovery time is determined as the target recovery time interval.

[0223] Step 340: Determine the risk level corresponding to the target recovery time interval as the user's second risk level. The second risk level is used to characterize the risk level of the user's postoperative recovery.

[0224] The second risk level refers to a qualitative label defined based on the recovery time interval, such as low risk, medium risk, and high risk, reflecting the possibility of a delay in the recovery process.

[0225] For example, based on the preset mapping relationship between recovery time intervals and risk levels, the risk level mapped to the target recovery time interval is determined as the user's second risk level.

[0226] Step 350: Generate a postoperative management strategy for the user based on the first risk level and the second risk level.

[0227] Postoperative management strategies can include personalized intervention plans developed by combining primary and secondary risk levels. These can specifically include monitoring recommendations, recovery management recommendations, and follow-up plans. For example, monitoring recommendations could include hourly vital sign monitoring and daily brain oxygen saturation assessment for high-risk patients, which is related to monitoring for complications relevant to the mechanisms of anesthesia. Recovery management recommendations could include early rehabilitation training. Follow-up plans could include daily follow-up for the first 3 days post-surgery for patients with high primary and / or secondary risk levels, and follow-up for 1 week post-surgery for patients with low primary and secondary risk levels.

[0228] For example, a mapping rule is established between risk level combinations and postoperative management strategies, and the postoperative management strategy mapped to the risk level combination of the first risk level and the second risk level is the postoperative management strategy for the user.

[0229] Therefore, a postoperative management strategy for a user is determined based on a risk level combination consisting of a first risk level that characterizes the user's risk level of complications and a second risk level that characterizes the user's postoperative recovery, thus achieving personalized and precise postoperative management for the user.

[0230] In some embodiments of this application, the above-mentioned method for determining postoperative recovery time may further include outputting an interpretability report, specifically including key influencing factor analysis, decision path visualization, similar case reference, and expert advice and knowledge mapping. The key influencing factor analysis may employ the SHAP value analysis method to quantify the contribution of each input feature to the final prediction result (such as intraoperative anesthesia method recommendation and postoperative recovery time), generating a feature importance ranking so that medical staff can intuitively understand which factors play a dominant role in the model decision. Decision tree path visualization technology clearly displays the complete logical chain from input parameters (such as patient age, surgical type, underlying diseases) to anesthesia method decision, and then to postoperative recovery time prediction, making the model decision derivation process transparent. Based on patient characteristics (such as intraoperative anesthesia method, complication risk, recovery time, etc.), a similarity algorithm is used to retrieve historical similar cases and provide reference comparisons. Expert clinical experience is integrated to form targeted recommendations, and expert knowledge mapping is used to compare the model decision with expert experience, marking the consistency and differences between the two.

[0231] In some embodiments of this application, the above-mentioned method for determining postoperative recovery time may further include establishing a model of the influence mechanism of anesthesia mode on postoperative recovery time, which may specifically include a direct influence model, an indirect influence model, and an interactive influence model. Subsequently, the above-mentioned key influencing factors analysis and decision tree path are determined based on the influence mechanism model.

[0232] The direct impact model can employ deep neural networks to model the direct effect of the anesthesia method itself on postoperative recovery time. This direct impact model quantifies this direct effect by learning the correlation between drug metabolism patterns, anesthesia depth, and recovery indicators. The indirect impact model can use a random forest algorithm to focus on the indirect effect of the anesthesia method on recovery time through mediating variables such as complications. Mediating variables for complications include: perioperative surgery-related complications such as intraoperative massive hemorrhage, vascular injury, and coronary artery obstruction requiring emergency intervention, or postoperative complications such as new-onset arrhythmias and paravalvular leak requiring reoperation; and perioperative anesthesia-related complications such as airway incontinence under non-general anesthesia, circulatory collapse after general anesthesia, respiratory failure (prolonged mechanical ventilation), stroke, and acute kidney injury. This indirect impact model quantifies the indirect effect by analyzing the chain relationship of anesthesia method → ​​complications → prolonged recovery time. This indirect impact model can be the aforementioned complication risk assessment model. The interaction impact model can use a gradient boosting algorithm to model the interaction between patient characteristics and anesthesia method on recovery time.

[0233] In some embodiments of this application, the above-mentioned method for determining postoperative recovery time may further include: establishing multi-center data collection standards to ensure the consistency and comparability of data from different medical institutions, providing high-quality training samples for the anesthesia impact mechanism model, and avoiding overfitting of the model to specific center data due to data bias; adopting a federated learning method to achieve collaborative training of multi-center models while protecting patient privacy, thus avoiding the privacy risks of cross-institutional data sharing and integrating diverse cases from multiple centers to improve the generalization ability of the anesthesia impact mechanism model; establishing model performance evaluation standards, including multi-dimensional indicators such as prediction accuracy, clinical usability, and expert satisfaction; and developing standardized procedures for model deployment and application to ensure consistent performance in different medical environments.

[0234] It should also be noted that the postoperative recovery time determination method provided in this application embodiment is specifically used for the postoperative recovery time of TAVR patients, and can also be extended to other cardiovascular interventional procedures for anesthesia decision-making and recovery time prediction scenarios, including percutaneous coronary intervention (PCI), percutaneous mitral valve clipping (MitraClip), and percutaneous pulmonary valve implantation (PPVI), etc. Specifically, at the model level, targeted adaptation can be achieved by adjusting the parameter weights and expert rules within the model.

[0235] Based on the postoperative recovery time determination method provided in the above embodiments, this application also provides specific implementation methods of the postoperative recovery time determination device. Please refer to the following embodiments.

[0236] See Figure 8 The postoperative recovery time determination device 40 provided in this application embodiment includes:

[0237] The acquisition module 410 is used to acquire user parameters; user parameters include physiological characteristic parameters that characterize the user's physiological state, disease risk level parameters that characterize the user's disease, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters;

[0238] The first determining module 420 is used to input user parameters into the intraoperative anesthesia method determining model and obtain the intraoperative anesthesia method output by the intraoperative anesthesia method determining model. The intraoperative anesthesia method includes non-general anesthesia method or general anesthesia method.

[0239] The second determining module 430 is used to determine the user's postoperative recovery time based on the intraoperative anesthesia method and user parameters.

[0240] Therefore, the acquisition module 410 can acquire multi-dimensional user parameters, specifically including physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters. These multi-dimensional user parameters comprehensively depict the factors affecting the user's postoperative recovery from multiple perspectives, such as the user's own physiological state, disease risk, surgical operation, anesthesia effects, and the medical team's level, avoiding the limitations of single-dimensional data. Using the first determination module 420, the model is determined through intraoperative anesthesia method determination. Based on the multi-dimensional user parameters, the specific intraoperative anesthesia method is first determined. Then, the second determination module 430 combines the multi-dimensional user parameters and the intraoperative anesthesia method to jointly determine the postoperative recovery time. This makes the determination process of postoperative recovery time more closely reflect the user's actual clinical scenario. Thus, by integrating more key variables affecting the user's postoperative recovery, the accuracy of the determined postoperative recovery time is effectively improved.

[0241] In some embodiments of this application, the intraoperative anesthesia method determination model includes an applicability assessment module, an applicability adjustment module, a complexity assessment module, and a decision module. The first determination module 420 is specifically used to: input user parameters into the applicability assessment module, the applicability adjustment module, and the complexity assessment module respectively, to obtain a first applicability score corresponding to the general anesthesia method and a second applicability score corresponding to the non-general anesthesia method output by the applicability assessment module, an applicability score adjustment factor output by the applicability adjustment module, and a complexity index output by the complexity assessment module; the first applicability score characterizes the applicability of the user using the general anesthesia method; the second applicability score characterizes the applicability of the user using the non-general anesthesia method; the applicability score adjustment factor characterizes the adjustment range of the applicability score corresponding to the general anesthesia method and the applicability score corresponding to the non-general anesthesia method; and the complexity index characterizes the complexity faced by the user in implementing the general anesthesia method and the non-general anesthesia method; input the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index into the decision module to obtain the intraoperative anesthesia method output by the decision module.

[0242] In some embodiments of this application, the first determining module 420 is specifically used to: input the first applicability score, the second applicability score, the applicability score adjustment factor, and the complexity index into the decision module to obtain the first comprehensive score corresponding to the general anesthesia method and the second comprehensive score corresponding to the non-general anesthesia method; and through the decision module, determine the anesthesia method corresponding to the larger value between the first comprehensive score and the second comprehensive score as the intraoperative anesthesia method output by the decision module.

[0243] In some embodiments of this application, the second determining module 430 is specifically used for: selecting target parameter types associated with the intraoperative anesthesia method from parameter types, including physiological characteristic parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters, and medical team capability evaluation parameters; selecting target parameters associated with the intraoperative anesthesia method from the target parameter set corresponding to the target parameter type; and inputting the target parameters into the postoperative recovery time determination model to obtain the postoperative recovery time output by the postoperative recovery time determination model.

[0244] In some embodiments of this application, the postoperative recovery time determination model includes a first postoperative recovery time determination model and a second postoperative recovery time determination model; the second determination module 430 is specifically used for: when the intraoperative anesthesia method is not general anesthesia, inputting the target parameter into the first postoperative recovery time determination model to obtain the postoperative recovery time output by the first postoperative recovery time determination model; when the intraoperative anesthesia method is general anesthesia, inputting the target parameter into the second postoperative recovery time determination model to obtain the postoperative recovery time output by the second postoperative recovery time determination model.

[0245] In some embodiments of this application, the first postoperative recovery time determination model includes a first immediate recovery time determination module, a first short-term recovery time determination module, and a first long-term recovery time determination module. The second determination module 430 is specifically used to: input target parameters into the first immediate recovery time determination module, the first short-term recovery time determination module, and the first long-term recovery time determination module respectively, to obtain the first immediate recovery time output by the first immediate recovery time determination module, the first short-term recovery time output by the first short-term recovery time determination module, and the first long-term recovery time output by the first long-term recovery time determination module; the first immediate recovery time characterizes the time required for a user to complete a first type of immediate recovery event within a first preset time window after surgery when using non-general anesthesia during surgery; the first short-term recovery time characterizes the time required for a user to complete a first type of short-term recovery event within a second preset time window after surgery when using non-general anesthesia during surgery; and the first immediate recovery time, the first short-term recovery time, and the first long-term recovery time are determined as the postoperative recovery time output by the first postoperative recovery time determination model.

[0246] In some embodiments of this application, the second postoperative recovery time determination model includes a second immediate recovery time determination module, a second short-term recovery time determination module, and a second long-term recovery time determination module. The second determination module 430 is specifically used to: input target parameters into the second immediate recovery time determination module, the second short-term recovery time determination module, and the second long-term recovery time determination module, respectively, to obtain the second immediate recovery time output by the second immediate recovery time determination module, the second short-term recovery time output by the second short-term recovery time determination module, and the second long-term recovery time output by the second long-term recovery time determination module; the second immediate recovery time characterizes the time required for a user to complete a second type of immediate recovery event within a second preset time window after surgery when using non-general anesthesia during surgery; the second short-term recovery time characterizes the time required for a user to complete a second type of short-term recovery event within a second preset time window after surgery when using non-general anesthesia during surgery; and the second immediate recovery time, the second short-term recovery time, and the second long-term recovery time are determined as the postoperative recovery time output by the second postoperative recovery time determination model.

[0247] In some embodiments of this application, the postoperative recovery time determination device 40 further includes: a first processing module, used to input the intraoperative anesthesia method and user parameters into a complication risk assessment model to obtain the complication type output by the complication risk assessment model and the risk probability of the user having the corresponding complication type; a second processing module, used to input the complication type, risk probability, and postoperative recovery time into an anesthesia influence determination model to obtain the indirect effect value output by the anesthesia influence determination model; the indirect effect value is used to characterize the degree of influence of the intraoperative anesthesia method on the postoperative recovery time through complications; and a third processing module, used to adjust the postoperative recovery time through the indirect effect value to obtain a target recovery time, the target recovery time being the user's postoperative recovery time under the influence of complications.

[0248] In some embodiments of this application, the postoperative recovery time determination device 40 further includes: a third determining module, configured to determine a target risk probability interval matching the risk probability from multiple preset risk probability intervals corresponding to the complication based on the user's risk probability of developing complications; a fourth determining module, configured to determine the risk level corresponding to the target risk probability interval as the user's first risk level, the first risk level being used to characterize the user's risk level of developing complications; a fifth determining module, configured to determine a target recovery time interval matching the target recovery time from multiple preset recovery time intervals based on the target recovery time; a sixth determining module, configured to determine the risk level corresponding to the target recovery time interval as the user's second risk level, the second risk level being used to characterize the user's postoperative recovery risk level; and a generating module, configured to generate a postoperative management strategy for the user based on the first risk level and the second risk level.

[0249] The various modules of the postoperative recovery time determination device 40 provided in this application embodiment can achieve... Figures 1 to 7 It provides functions for each step of the method for determining postoperative recovery time and achieves the corresponding technical effects. For the sake of brevity, it will not be described in detail here.

[0250] Figure 9 The diagram shows a schematic representation of the hardware structure of an electronic device provided in some embodiments of this application.

[0251] The electronic device may include a processor 501 and a memory 502 storing computer program instructions.

[0252] Specifically, the processor 501 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0253] Memory 502 may include mass storage for data or instructions. For example, and not limitingly, memory 502 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 502 may include removable or non-removable (or fixed) media. Where appropriate, memory 502 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 502 is non-volatile solid-state memory.

[0254] In a particular embodiment, memory 502 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, typically, memory 502 includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the postoperative recovery time determination method according to the above embodiments of this application.

[0255] The processor 501 reads and executes computer program instructions stored in the memory 502 to implement any of the postoperative recovery time determination methods in the above embodiments.

[0256] In one example, the electronic device may also include a communication interface 503 and a bus 510. Wherein, as... Figure 5 As shown, the processor 501, memory 502, and communication interface 503 are connected through bus 510 and complete communication with each other.

[0257] The communication interface 503 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0258] Bus 510 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 510 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0259] This electronic device can execute the postoperative recovery time determination method in the embodiments of this application, thereby achieving a combination of Figures 1 to 8 The method and apparatus for determining postoperative recovery time are described.

[0260] Furthermore, in conjunction with the postoperative recovery time determination method in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the postoperative recovery time determination methods in the above embodiments. Examples of computer-readable storage media include non-transitory computer-readable storage media, such as portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, etc.

[0261] Furthermore, in conjunction with the postoperative recovery time determination method in the above embodiments, this application embodiment can provide a computer program product to implement it. This program product is stored in a storage medium and may specifically include a computer program or instructions. When executed by a processor, the computer program or instructions implement any of the postoperative recovery time determination methods in the above embodiments. This program product is executed by at least one processor to implement the various processes of the above-described postoperative recovery time determination method embodiments and achieves the same technical effects; to avoid repetition, further details are omitted here.

[0262] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0263] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0264] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0265] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0266] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A postoperative recovery time determination method characterized by, The method comprises the following steps: obtaining user parameters of a user; the user parameters comprise physiological feature parameters for representing physiological states of the user, disease risk level parameters for representing diseases suffered by the user, surgery-related parameters, anesthesia-related parameters and medical team capability evaluation parameters; inputting the user parameters into an intraoperative anesthesia mode determination model to obtain an intraoperative anesthesia mode output by the intraoperative anesthesia mode determination model; the intraoperative anesthesia mode comprises a non-general anesthesia mode or a general anesthesia mode; determining a postoperative recovery time of the user according to the intraoperative anesthesia mode and the user parameters.

2. The method of claim 1, wherein, The intraoperative anesthesia mode determination model comprises a applicability evaluation module, an applicability adjustment module, a complexity evaluation module and a decision module; The step of inputting the user parameters into the intraoperative anesthesia mode determination model to obtain the intraoperative anesthesia mode output by the intraoperative anesthesia mode determination model comprises the following steps: inputting the user parameters into the applicability evaluation module, the applicability adjustment module and the complexity evaluation module respectively to obtain a first applicability score corresponding to the general anesthesia mode and a second applicability score corresponding to the non-general anesthesia mode output by the applicability evaluation module, an applicability score adjustment factor output by the applicability adjustment module and a complexity index output by the complexity evaluation module; the first applicability score is used to represent the applicability of the general anesthesia mode to the user; the second applicability score is used to represent the applicability of the non-general anesthesia mode to the user; the applicability score adjustment factor is used to represent the adjustment range of the applicability scores corresponding to the general anesthesia mode and the non-general anesthesia mode; and the complexity index is used to represent the complexity degree faced by the user when the general anesthesia mode and the non-general anesthesia mode are implemented; inputting the first applicability score, the second applicability score, the applicability score adjustment factor and the complexity index into the decision module to obtain the intraoperative anesthesia mode output by the decision module.

3. The method of claim 2, wherein, The step of inputting the first applicability score, the second applicability score, the applicability score adjustment factor and the complexity index into the decision module to obtain the intraoperative anesthesia mode output by the decision module comprises the following steps: inputting the first applicability score, the second applicability score, the applicability score adjustment factor and the complexity index into the decision module to obtain a first comprehensive score corresponding to the general anesthesia mode and a second comprehensive score corresponding to the non-general anesthesia mode; determining, by the decision module, the anesthesia mode corresponding to the larger value between the first comprehensive score and the second comprehensive score as the intraoperative anesthesia mode output by the decision module.

4. The method of claim 1, wherein, The step of determining the postoperative recovery time of the user according to the intraoperative anesthesia mode and the user parameters comprises the following steps: screening, according to the intraoperative anesthesia mode, a target parameter type associated with the intraoperative anesthesia mode from parameter types; the parameter types comprise physiological feature parameters, disease risk level parameters, surgery-related parameters, anesthesia-related parameters and medical team capability evaluation parameters; According to the intraoperative anesthesia mode, a target parameter associated with the intraoperative anesthesia mode is screened from a target parameter set corresponding to a target parameter type; The target parameter is input into a postoperative recovery time determination model, and the postoperative recovery time output by the postoperative recovery time determination model is obtained.

5. The method of claim 4, wherein, The postoperative recovery time determination model includes a first postoperative recovery time determination model and a second postoperative recovery time determination model; the target parameter is input into the postoperative recovery time determination model, and the postoperative recovery time output by the postoperative recovery time determination model is obtained, including: In a case where the intraoperative anesthesia mode is the non-general anesthesia mode, the target parameter is input into the first postoperative recovery time determination model, and the postoperative recovery time output by the first postoperative recovery time determination model is obtained; In a case where the intraoperative anesthesia mode is the general anesthesia mode, the target parameter is input into the second postoperative recovery time determination model, and the postoperative recovery time output by the second postoperative recovery time determination model is obtained.

6. The method of claim 5, wherein, The first postoperative recovery time determination model includes a first immediate recovery time determination module, a first short-term recovery time determination module, and a first long-term recovery time determination module; The target parameter is input into the first postoperative recovery time determination model, and the postoperative recovery time output by the first postoperative recovery time determination model is obtained, including: The target parameter is input into the first immediate recovery time determination module, the first short-term recovery time determination module, and the first long-term recovery time determination module respectively, and a first immediate recovery time output by the first immediate recovery time determination module, a first short-term recovery time output by the first short-term recovery time determination module, and a first long-term recovery time output by the first long-term recovery time determination module are obtained; the first immediate recovery time is used to represent a time required by the user to complete a first type of immediate recovery event within a first preset time window after surgery when the non-general anesthesia mode is adopted during surgery; the first short-term recovery time is used to represent a time required by the user to complete a first type of short-term recovery event within a second preset time window after surgery when the non-general anesthesia mode is adopted during surgery; and the first long-term recovery time is used to represent a time required by the user to complete a first type of long-term recovery event within a third preset time window after surgery when the non-general anesthesia mode is adopted during surgery; The first immediate recovery time, the first short-term recovery time, and the first long-term recovery time are determined as the postoperative recovery time output by the first postoperative recovery time determination model.

7. The method of claim 5, wherein, The second postoperative recovery time determination model includes a second immediate recovery time determination module, a second short-term recovery time determination module, and a second long-term recovery time determination module; The target parameter is input into the second postoperative recovery time determination model, and the postoperative recovery time output by the second postoperative recovery time determination model is obtained, including: inputting the target parameters into the second immediate recovery time determination module, the second short-term recovery time determination module and the second long-term recovery time determination module, obtaining a second immediate recovery time output by the second immediate recovery time determination module, a second short-term recovery time output by the second short-term recovery time determination module and a second long-term recovery time output by the second long-term recovery time determination module; the second immediate recovery time is used to represent a time required by the user to complete a second type of immediate recovery event within a second preset time window after surgery when the user adopts the general anesthesia method during surgery; the second short-term recovery time is used to represent a time required by the user to complete a second type of short-term recovery event within a second preset time window after surgery when the user adopts the general anesthesia method during surgery; and the second long-term recovery time is used to represent a time required by the user to complete a second type of long-term recovery event within a third preset time window after surgery when the user adopts the general anesthesia method during surgery; determining the second immediate recovery time, the second short-term recovery time and the second long-term recovery time as the postoperative recovery time output by the second postoperative recovery time determination model.

8. The method of claim 1, wherein, The method further comprises: inputting the intraoperative anesthesia method and the user parameters into a complication risk assessment model to obtain a complication type and a risk probability of the user suffering from a complication corresponding to the complication type output by the complication risk assessment model; inputting the complication type, the risk probability and the postoperative recovery time into an anesthesia influence determination model to obtain an indirect effect value output by the anesthesia influence determination model; the indirect effect value is used to represent an influence degree of the intraoperative anesthesia method on the postoperative recovery time through the complication; adjusting the postoperative recovery time through the indirect effect value to obtain a target recovery time, which is a postoperative recovery time of the user under the influence of the complication.

9. The method of claim 8, wherein, The method further comprises: determining a target risk probability interval matching the risk probability from a plurality of preset risk probability intervals corresponding to the complication according to the risk probability of the user suffering from the complication; determining a first risk level of the user as a risk level corresponding to the target risk probability interval, the first risk level being used to represent a risk level of the user suffering from the complication; determining a target recovery time interval matching the target recovery time from a plurality of preset recovery time intervals according to the target recovery time; determining a second risk level of the user as a risk level corresponding to the target recovery time interval, the second risk level being used to represent a risk level of the user recovering after surgery; generating a postoperative management strategy for the user according to the first risk level and the second risk level.

10. A postoperative recovery time determination apparatus characterized by comprising: The device comprises: an acquisition module configured to acquire user parameters of a user; the user parameters comprise physiological feature parameters used to represent physiological states of the user, disease risk level parameters used to represent diseases suffered by the user, surgery-related parameters, anesthesia-related parameters and medical team capability evaluation parameters; The first determining module is configured to input the user parameter into an intraoperative anesthesia mode determining model, and obtain an intraoperative anesthesia mode output by the intraoperative anesthesia mode determining model, the intraoperative anesthesia mode including a non-general anesthesia mode or a general anesthesia mode. The second determining module is configured to determine the postoperative recovery time of the user according to the intraoperative anesthesia mode and the user parameter.

11. An electronic device, comprising: The device comprises a processor and a memory storing computer program instructions; The processor executes the computer program instructions to implement the postoperative recovery time determining method in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer program instructions stored on the computer readable storage medium are executed by the processor to implement the postoperative recovery time determining method in any one of claims 1-9.

13. A computer program product, characterised in that, The instructions in the computer program product are executed by the processor of the electronic device to enable the electronic device to implement the postoperative recovery time determining method in any one of claims 1-9.

Images