Preoperative visit assessment system for anesthesia risk

Abstract

The application discloses an anesthetic risk preoperative visit evaluation system and relates to the technical field of anesthetic risk evaluation.The application retains existing scoring methods such as CAT index and MET activity equivalent, introduces an optimization mechanism based on data classification analysis and weight self-adaptive calculation, makes the scoring process more scientific and objective, and has individual characteristics, realizes error correction and dynamic updating of the scoring result by normalizing the comprehensive anesthetic risk value and the initial score and taking the average value, and significantly reduces the risk of clinical misjudgment caused by scoring deviation.In conclusion, the application introduces classification risk evaluation, a double weight mechanism and a dynamic fusion strategy, constructs a more scientific and intelligent anesthetic risk scoring optimization system, and provides strong support for individualized formulation of preoperative anesthetic schemes.

Description

Technical Field

[0001] This invention relates to the field of anesthesia risk assessment technology, and more particularly to a preoperative anesthesia risk assessment system. Background Technology

[0002] With the continuous development of artificial intelligence technology in the medical field, intelligent patient visits, as an important component of digital healthcare, have demonstrated significant advantages in preoperative assessment. Intelligent patient visits refer to the use of artificial intelligence systems, combined with patients' basic personal information data, including electronic medical records and past medical history, to assist doctors in completing the preoperative visit process through intelligent question answering, automatic analysis, and risk identification. Its core lies in utilizing technologies such as natural language processing and risk prediction models to achieve rapid identification and preliminary assessment of patients' health conditions, thereby improving the efficiency and standardization of preoperative visits.

[0003] In preoperative anesthesia assessments, anesthesia risk scoring is a crucial basis for evaluating perioperative safety and developing anesthesia protocols. Currently, commonly used anesthesia risk scoring methods are mostly based on single indicators, such as the CAT score for patients with chronic obstructive pulmonary disease (COPD), MET activity equivalent assessment, coagulation function status, and degree of dyspnea. These methods typically combine multiple independent indicators linearly to form a comprehensive score reflecting the patient's overall anesthesia risk level. However, existing scoring systems are mostly fixed and static, meaning their scoring rules remain unchanged once established. This is primarily because existing scoring systems are often based on early clinical research or expert consensus, and their scoring standards are used extensively and difficult to update in a timely manner based on new clinical evidence. This often leads to discrepancies between the scoring results and the actual risk level in practical applications. For example, when dealing with patients with multiple underlying diseases, advanced age, or marginal organ function compensation, static scoring models cannot effectively identify the changing trends of their potential risk factors, thus overestimating or underestimating the patient's anesthesia tolerance. Therefore, if the existing static scoring system is still used, lacking dynamic optimization, the deviation in the scoring results can directly affect the choice of anesthesia plan. If the score is too low, it may lead anesthesiologists to underestimate the patient's risk and choose a relatively aggressive anesthesia method, increasing the incidence of intraoperative or postoperative complications; while if the score is too high, it may cause the patient to be unreasonably excluded from certain tolerable surgeries, delaying treatment and affecting prognosis. Therefore, there is an urgent need for a technical solution of a preoperative anesthesia risk assessment system. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a preoperative anesthesia risk assessment system, specifically comprising the following modules:

[0005] Data acquisition module: used by patients to log in to the system by scanning a QR code and fill in their basic personal information;

[0006] Data analysis module: connected to the data acquisition module, the data analysis module includes a final weight determination subunit, which is used to analyze the patient's personal basic information data according to the final weight determination subunit to obtain the final patient's anesthesia risk score;

[0007] Final weight determination subunit: Used to determine the final weight of each data category based on the patient's anesthesia risk value under each data category. The final weight determination subunit includes the following interfaces;

[0008] First weight calculation interface: used to determine the first weight of each data category based on the ratio of the patient's anesthesia risk value under each data category to the sum of the anesthesia risk values ​​of patients under all data categories;

[0009] The second weight determination interface is used to analyze the correlation between the anesthesia risk values ​​of patients under each data category and determine the second weight of each data category based on the analysis results.

[0010] Final weight calculation interface: used to take the average of the sum of the first weight and the second weight of each data class as the final weight of each data class;

[0011] Index Calculation Unit: Used to calculate the patient's Chronic Obstructive Pulmonary Disease (COPD) CAT Index, Coagulation Disorder Propensity Index, MET Activity Equivalent Index, and Dyspnea Index based on the patient's basic personal information data;

[0012] Initial anesthesia risk score calculation unit: used to comprehensively calculate the patient's initial anesthesia risk score by combining the chronic obstructive pulmonary disease CAT index, coagulation disorder tendency index, MET activity equivalent index, and dyspnea index;

[0013] Scoring optimization unit: Used to optimize the patient's initial anesthesia risk score to obtain the final anesthesia risk score for the patient.

[0014] Data classification subunit: used to classify patients' basic personal information data according to data type;

[0015] Risk value calculation subunit: used to calculate the anesthesia risk value for patients under each data category;

[0016] Clinical parameter extraction interface: used to extract clinical parameters of patients under each data category;

[0017] Clinical parameter normalization interface: used to normalize each clinical parameter of a patient under each type of data;

[0018] Partial correlation coefficient calculation interface: used to construct the function mapping relationship between each clinical parameter within each data class based on each clinical parameter of the patient under each normalized data class, and calculate the partial correlation coefficient between each clinical parameter based on the function mapping relationship;

[0019] Contribution factor acquisition interface: used to determine the direction and intensity of the influence of each clinical parameter on anesthesia risk based on the partial correlation coefficient, and to generate the contribution factor of each clinical parameter for each patient under each type of data;

[0020] Original anesthesia risk score acquisition interface: This is used to perform a product operation on each clinical parameter contribution factor of the patient under each data category and each normalized clinical parameter of the patient under each data category to obtain the original anesthesia risk score of the patient under each data category.

[0021] Anesthesia Risk Value Acquisition Interface: This interface is used to perform a non-linear transformation on the original anesthesia risk scores to obtain the anesthesia risk value for each type of data.

[0022] Final weight determination subunit: used to determine the final weight of each data category based on the patient's anesthesia risk value for each data category;

[0023] The anesthesia risk value matrix construction sub-interface is used to construct an anesthesia risk vector set consisting of the anesthesia risk values ​​of patients under each data category, and to form an anesthesia risk value matrix based on the anesthesia risk vector set.

[0024] The correlation coefficient matrix generation sub-interface is used to calculate the Pearson correlation coefficient between each type of data based on the anesthesia risk value matrix and generate the correlation coefficient matrix.

[0025] Eigenvalue decomposition sub-interface: Used to perform eigenvalue decomposition on the correlation coefficient matrix and extract the eigenvalues ​​and corresponding eigenvectors from the correlation coefficient matrix;

[0026] Principal Component Factor Extraction Sub-interface: Used to sort eigenvalues ​​in descending order, calculate the variance contribution rate and cumulative variance contribution rate of each eigenvalue, and select the eigenvectors corresponding to the first k eigenvalues ​​whose cumulative variance contribution rate reaches a preset threshold as principal component factors;

[0027] Loading coefficient extraction sub-interface: used to extract loading coefficients from principal component factors and map the loading coefficients back to the data categories that constitute the anesthesia risk value matrix, and identify the data categories that have a significant impact on the principal components;

[0028] The second weight calculation sub-interface is used to calculate the second weight of each data class based on the variance contribution rate of the principal component factors and the corresponding loading coefficients.

[0029] The comprehensive anesthesia risk value calculation subunit is used to combine the final weight of each type of data with the anesthesia risk value of the patient under each type of data to obtain the patient's comprehensive anesthesia risk value.

[0030] The Comprehensive Anesthesia Risk Score Acquisition Subunit is used to normalize the patient's comprehensive anesthesia risk value to obtain the patient's comprehensive anesthesia risk score.

[0031] Optimized score acquisition sub-unit: This unit combines the patient's comprehensive anesthesia risk score with the patient's initial anesthesia risk score and takes the average to obtain the final anesthesia risk score for the patient.

[0032] Anesthesia planning module: Connected to the data analysis module, it is used to develop anesthesia plans based on the final patient's anesthesia risk score.

[0033] The embodiments of the present invention have the following technical effects:

[0034] The core of this invention lies in the systematic optimization of existing anesthesia risk scoring mechanisms to address the inaccuracy of traditional scoring systems due to their static nature and lack of dynamic adjustment capabilities. While retaining existing scoring methods such as the CAT index and MET activity equivalent, this invention introduces an optimization mechanism based on data classification analysis and adaptive weight calculation, making the scoring process more scientific, objective, and individualized.

[0035] This invention categorizes patients' basic information data, calculates the anesthesia risk value for each category, determines the final weight of each data type, and then combines the comprehensive anesthesia risk value with the initial score for dynamic correction of the existing scoring results. This mechanism breaks through the limitations of the uniform standard in traditional scoring models, allowing scores to be flexibly adjusted according to the actual contribution of different data categories, thus improving the personalization of the scoring.

[0036] Based on this, this invention constructs a dual-weight assessment mechanism consisting of two types of weights, and forms the final weight by fusing the two types of weights, thereby improving the adaptability and stability of the scoring system. The first weight is determined based on the proportion of risk value in the overall data for each data type, reflecting the relative influence of that data type in the current patient risk assessment. The second weight is calculated based on the correlation between different data types, by extracting principal component factors through principal component analysis and combining loading coefficients and variance contribution rates. Its function is to identify the correlation between different data categories in terms of risk structure, reflecting the structural position of a certain data type in the entire scoring system.

[0037] The final weight, formed by merging the first and second weights, not only reflects the risk contribution of each data category but also considers the interaction between this data category and other data categories. This allows the scoring system to reflect individual differences while also taking into account the overall structural characteristics. Compared to the traditional approach of directly and linearly combining multiple indicators, this fusion mechanism is more logical and interpretable, effectively improving the accuracy and robustness of the scoring results.

[0038] Ultimately, by normalizing the comprehensive anesthesia risk value and the initial score and taking the average, error correction and dynamic updating of the scoring results were achieved, significantly reducing the risk of clinical misjudgment due to scoring bias. In summary, this invention, by introducing categorical risk assessment, a dual-weighting mechanism, and a dynamic fusion strategy, constructs a more scientific and intelligent anesthesia risk scoring optimization system, providing strong support for the personalized development of preoperative anesthesia plans. Attached Figure Description

[0039] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0040] Figure 1 This is a framework diagram of the preoperative visit and assessment system for anesthesia risk provided in an embodiment of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0042] Example 1: As Figure 1 As shown, the preoperative anesthesia risk assessment system provided by this invention includes the following modules:

[0043] Data acquisition module: used by patients to log in to the system by scanning a QR code and fill in their basic personal information;

[0044] It's worth noting that the QR code generation process involves the doctor first logging into the pre-anesthesia visit system using their unique account. This system assigns a unique account and password to each registered doctor to ensure information security and privacy. Then, on the system homepage, the doctor can find the "Generate Unique QR Code" function. Clicking this function generates a unique QR code based on the doctor's identity information, such as doctor ID and department. The QR code contains not only the doctor's basic information but also information related to the current surgery or visit, such as the type of surgery and the estimated time, so that the system can accurately link the patient and doctor and their surgical schedule. The doctor can then share the generated QR code with the patient in various ways, such as via email, SMS, the hospital's internal system, or by printing it out directly. The validity period of the QR code can be set as needed, typically one-time or valid for a certain period to ensure information security.

[0045] After scanning the QR code shared by the doctor, the system guides the patient to complete a series of basic information entries. This includes, but is not limited to, the following: patient's name, gender, date of birth, contact information, ID number; current medical history, such as the patient's major health problems (e.g., whether they have chronic diseases like hypertension or diabetes), and whether they have recently received other treatments or surgeries; systemic medical history, such as past major illnesses like heart disease, respiratory diseases, or kidney disease, previous surgical experiences, especially any anesthesia-related adverse reactions; personal life history, such as smoking history, alcohol consumption habits, and daily activity levels. This information helps assess the patient's overall health and anesthesia tolerance; and preliminary physical examination results, such as key indicators from the most recent physical examination, such as blood pressure, heart rate, and blood oxygen saturation. Through the detailed QR code generation process and the example of the basic information entered by the patient, the pre-anesthesia visit system effectively collects relevant patient information and provides a solid data foundation for subsequent anesthesia risk assessment. This process not only improves the efficiency of pre-operative visits but also enhances the accuracy of assessment results, thereby better ensuring patient safety.

[0046] Data analysis module: connected to the data acquisition module, the data analysis module includes a final weight determination subunit, which is used to analyze the patient's personal basic information data according to the final weight determination subunit to obtain the final patient's anesthesia risk score;

[0047] Final weight determination subunit: used to determine the final weight of each data category based on the patient's anesthesia risk value for each data category;

[0048] First weight calculation interface: used to determine the first weight of each data category based on the ratio of the patient's anesthesia risk value under each data category to the sum of the anesthesia risk values ​​of patients under all data categories;

[0049] The second weight determination interface is used to analyze the correlation between the anesthesia risk values ​​of patients under each data category and determine the second weight of each data category based on the analysis results.

[0050] Final weight calculation interface: used to take the average of the sum of the first weight and the second weight of each data class as the final weight of each data class;

[0051] The system's design, from the first weight calculation interface to the final weight calculation interface, revolves around the generation mechanism of the "final weight for each data category." It proposes a dual-weight fusion strategy: a mean-based approach using the first weight (based on the proportion of risk values) and the second weight (based on correlation analysis). This allows for dynamic identification and reasonable allocation of importance across different data categories. The key to this technical solution lies in capturing local risk contribution and global variable structural characteristics through two different weight calculation methods, thereby fusing them to obtain a more representative and robust final weight. The first weight calculates the proportion of anesthesia risk value within the overall risk value for each data category, reflecting the relative importance of that data category in the current sample. This method is intuitive and concise, quickly identifying data categories that contribute significantly to overall risk. The second weight, based on principal component analysis, extracts the main factors driving risk changes and combines variance contribution rate and loading coefficients to further determine which data categories play a major role in these principal components, thus obtaining a more statistically significant weight allocation. Finally, the system averages the two weights to obtain the final weight for each data category. This design fully considers the complementarity of the two weights: the first weight emphasizes the contribution of local risk, while the second weight focuses on the structural relationship between global variables. The combination of the two not only avoids the potential bias of a single weight but also enhances the algorithm's adaptability to different patient groups. Furthermore, this weight generation mechanism has significant advantages such as de-subjectification, dynamic adjustment, clear structure, and reliable results. It can automatically adjust the weight distribution according to differences in patient groups, improving the flexibility and generalization ability of the scoring. Therefore, the interface from the first weight calculation to the final weight calculation provides a scientific, flexible, and highly interpretable weight generation mechanism for the entire anesthesia risk assessment system. It serves as a crucial bridge connecting local risk calculation and global risk fusion, providing key support for the subsequent generation of comprehensive scores.

[0052] It is worth noting that, in order to improve the clinical applicability of the scoring optimization mechanism in this application, this application introduces optimization judgment conditions based on the patient's physical strength index, which is a quantitative indicator used to comprehensively assess the patient's overall physical functional status. The higher the index, the stronger the patient's physical condition and the stronger their tolerance to anesthesia and surgery; conversely, it suggests that the patient may have underlying diseases, functional decline, etc., requiring more refined anesthesia risk assessment and management.

[0053] To calculate the Body Strength Index, the following basic patient information data is required, including age. ,gender That is, binary data, BMI, and smoking status. Both are binary data and have activity capabilities. Number of comorbidities Lung function indicators hemoglobin concentration ,blood pressure The patient's physical strength index was then calculated using the following formula:

[0054] ;

[0055] in, The patient's physical strength index;

[0056] It is worth noting that the above data must be normalized before it can be included in the formula calculation to ensure that the index value is stable in the 0-1 range.

[0057] Then, based on the physical strength index of all patients in the historical data, its mean was calculated. with standard deviation Therefore, a threshold for determining whether to optimize or not is set, and the specific formula is as follows:

[0058]

[0059] The judgment was then executed, and the specific judgment logic is as follows:

[0060] If the patient's current physical strength index is less than the judgment threshold, then the subsequent scoring optimization step will be executed;

[0061] If the current patient's physical strength index is greater than or equal to the decision threshold, then skip the score optimization and use the initial score directly;

[0062] The reason for needing to optimize the scoring when a patient's physical strength index (PMI) is that a PMI below the threshold indicates weak overall physical function, which may suggest the following: multiple underlying diseases (e.g., numerous comorbidities); poor cardiopulmonary function (e.g., low FEV1, low MET); poor nutritional status (e.g., low BMI, low hemoglobin); decreased tolerance to anesthesia and surgery; and a higher risk of intraoperative or postoperative complications. In these patients, traditional scoring methods (such as CAT, MET, etc.) may underestimate the true risk due to the static and uniform nature of the scoring mechanism, leading to inappropriate anesthesia planning and increased intraoperative risks. Therefore, introducing a scoring optimization mechanism, through data classification analysis, adaptive weight adjustment, and multi-source scoring fusion, can more precisely identify individual patient risk characteristics, dynamically adjust scoring weights to highlight the impact of high-risk data categories, and improve the personalization and accuracy of scoring results, thereby providing safer and more reasonable anesthesia recommendations for clinicians.

[0063] The reason why score optimization is unnecessary when a patient's physical strength index is greater than or equal to the judgment threshold is that a physical strength index above or equal to the threshold usually indicates a good physical condition, characterized by: fewer underlying diseases, good cardiopulmonary function, good nutritional status, and strong tolerance to anesthesia and surgery. For these patients, traditional scoring methods (such as CAT and MET) already have high accuracy and applicability, and the scoring results are basically consistent with the actual risk level. While introducing a score optimization mechanism can provide more detailed risk analysis, it will bring the following problems: the scoring results fluctuate less, the optimized score is not significantly different from the initial score, and unnecessary complexity is easily introduced, affecting the doctor's understanding and trust in the scoring results. Therefore, for patients in good physical condition, a standard scoring mechanism can meet clinical needs, saving resources and improving efficiency and operability.

[0064] Therefore, by introducing the patient's physical strength index as the criterion for determining whether to optimize the scoring, the entire scoring system gains the ability to automatically identify the patient's risk level and dynamically adjust the scoring strategy, thus achieving a "personalized" scoring mechanism and enhancing the system's intelligence. Furthermore, this decision condition enables "layered processing" of the scoring process: high-risk patients receive refined assessments, while low-risk patients undergo simplified processing, avoiding unnecessary waste of computational resources and improving system efficiency. Simultaneously, by introducing the patient's physical strength index, the scoring system not only provides a numerical anesthesia risk score but also offers more clinically relevant decision support based on the patient's physical strength status, making the scoring results more interpretable and credible. In summary, the introduction of a scoring optimization decision condition based on the physical strength index in this invention not only improves the intelligence, adaptability, and efficiency of the entire scoring system but also provides a more scientific, accurate, and personalized anesthesia risk assessment mechanism for clinical practice. This mechanism allows patients with weaker physical conditions to receive more comprehensive and refined risk assessments, while sparing patients with good physical conditions from unnecessary complex assessments, improving efficiency. The entire scoring system maintains consistency and stability while possessing dynamic response and individualized processing capabilities.

[0065] Index Calculation Unit: Used to calculate the patient's Chronic Obstructive Pulmonary Disease (COPD) CAT Index, Coagulation Disorder Propensity Index, MET Activity Equivalent Index, and Dyspnea Index based on the patient's basic personal information data;

[0066] It is worth noting that the calculation process for the patient's Chronic Obstructive Pulmonary Disease (COPD) CAT Index, Coagulation Disorder Propensity Index, MET Activity Equivalent Index, and Dyspnea Index is as follows:

[0067] The calculation process of the patient's chronic obstructive pulmonary disease (CAT) index includes the following steps: First, the patient fills out a standard version of the "CAT Questionnaire", which has 8 items, each item is scored from 0 to 5, and the total score is 40 points. Then, the system automatically reads the original score value of each item.

[0068] The original CAT score, which has a total of 40 points, is then converted into a continuous value between 0 and 1 to fit the subsequent comprehensive scoring system.

[0069] The calculation process for a patient's coagulation propensity index includes extracting the following laboratory indicators from the test reports uploaded by the patient or entered by the doctor: prothrombin time (PT), in seconds; activated partial thromboplastin time (APTT), in seconds; fibrinogen concentration (FIB), in g / L; and then normalizing the above indicators based on the normal reference range.

[0070] If the actual value is below the lower limit, the normalized value is 0; if it is above the upper limit, the normalized value is 1.

[0071] Next, the arithmetic mean of the three normalized indicators is calculated to obtain the patient's coagulation disorder tendency index;

[0072] The calculation process for a patient's MET activity equivalent index includes obtaining the following structured questions completed by the patient in the system, with all answers being objective options: average weekly walking minutes, average weekly minutes of moderate-intensity exercise such as brisk walking or cycling, and average weekly minutes of high-intensity exercise such as running or skipping rope; then, a standard MET value conversion table is used:

[0073]

[0074] Then, calculate the daily average MET-minutes based on the standard MET value conversion table:

[0075]

[0076] in, , , The weekly cumulative time (in minutes) for walking, moderate-intensity exercise, and high-intensity exercise, respectively.

[0077] Finally, the daily average MET-minutes were standardized to obtain the patient's MET activity equivalent index.

[0078] The calculation process for the patient's dyspnea index includes the following steps: First, the patient fills out the standard version of the mMRC Dyspnea Scale questionnaire, which has 5 options (0-4 points), each corresponding to a different degree of dyspnea symptoms. The score is used directly as the baseline value and normalized to the [0,1] interval to obtain the corresponding dyspnea index for the patient.

[0079] Initial anesthesia risk score calculation unit: used to comprehensively calculate the patient's initial anesthesia risk score by combining the chronic obstructive pulmonary disease CAT index, coagulation disorder tendency index, MET activity equivalent index, and dyspnea index;

[0080] Scoring optimization unit: used to optimize the patient's initial anesthesia risk score to obtain the final anesthesia risk score;

[0081] The core technology of the data classification subunit to the optimized scoring acquisition subunit lies in dividing the patient's basic information data into categories, calculating the anesthesia risk value for each category, and determining the final weight of each category, thereby achieving dynamic adjustment and individualized optimization of the scoring process. By introducing a data classification analysis mechanism, the system can structure the originally mixed risk information, allowing risk factors of different dimensions to be calculated independently, avoiding information interference and improving the accuracy of risk identification. Based on this, the system introduces a dual-weight mechanism: the first weight reflects the relative contribution of each type of data in the current patient risk assessment, while the second weight identifies its structural position in the overall scoring system based on the correlation between data categories. The fusion of these two weights not only reflects the risk contribution of each data category itself but also considers its interaction with other data, ensuring that the scoring system reflects both individual differences and overall structural characteristics. Finally, the system combines the anesthesia risk value of each data category with the final weight to generate the patient's comprehensive anesthesia risk value, maps it to a standard range through normalization, and finally performs a weighted average with the initial score to form the final anesthesia risk score. This optimization process effectively smooths out fluctuations in scoring output, enhances the stability and interpretability of results, and significantly reduces the risk of clinical misjudgment due to scoring bias. Therefore, the data classification subunit to the optimized scoring acquisition subunit provides key technical support for achieving dynamic correction and individualized assessment of scoring results, and is an important component of this invention's construction of a scientific and intelligent anesthesia risk scoring system.

[0082] Data classification subunit: used to classify patients' basic personal information data according to data type;

[0083] It is worth noting that the categorized data includes basic demographic information, respiratory system-related data, circulatory and coagulation system-related data, exercise tolerance and metabolic function-related data, and past surgical and anesthesia history data. Furthermore, the basic demographic information category includes age, sex, height, weight, BMI, etc.; the respiratory system-related data category includes whether the patient has chronic obstructive pulmonary disease (COPD): yes / no; CAT score: 0-40 points; whether the patient smokes: yes / no; daily cigarette consumption: cigarettes / day; history of asthma: yes / no; FEV1 / FVC ratio: pulmonary function test result; peak expiratory flow rate: unit L / min; nighttime cough frequency: none / occasionally / frequently / every night, etc.; the circulatory and coagulation system-related data category includes whether the patient has a history of hypertension: yes / no; systolic blood pressure: unit mmHg; diastolic blood pressure: unit mmHg; heart rate: Units: bpm; History of coronary artery disease: Yes / No; Taking anticoagulants: Yes / No; Prothrombin time: Seconds; Activated partial thromboplastin time: Seconds; Fibrinogen concentration: G / L, etc.; Exercise tolerance and metabolic function related data include MET activity equivalent: MET; Diabetes: Yes / No; Fasting blood glucose: mmol / L; Glycated hemoglobin: %; Chronic kidney disease: Yes / No; Creatinine: μmol / L; eGFR (estimated glomerular filtration rate): mL / min / 1.73m²; Past surgical and anesthesia history data include: History of anesthesia allergy: Yes / No; Allergic drug name: Text input; Family history of malignant hyperthermia: Yes / No; History of serious post-anesthesia complications: Yes / No; History of general anesthesia: Yes / No; History of intubation difficulties: Yes / No, etc.

[0084] Risk value calculation subunit: used to calculate the anesthesia risk value for patients under each data category;

[0085] Clinical parameter extraction interface: used to extract clinical parameters of patients under each data category;

[0086] It is worth noting that when extracting clinical parameters for patients under each data category, the specific clinical parameters for each category are first extracted based on the data type classification results. Taking respiratory system-related data as an example, it may include the following six normalized clinical parameters, as detailed in the following table of clinical parameters for respiratory system-related data:

[0087]

[0088] Clinical parameter normalization interface: used to normalize each clinical parameter of a patient under each type of data;

[0089] It is worth noting that the Min-Max normalization method is used for continuous variables, the original value (0 or 1) is used directly for binary variables, and the range [0,1] is mapped to the ordinal categorical variable. For example, [0,4] is mapped to [0.0,0.25,0.5,0.75,1.0].

[0090] Partial correlation coefficient calculation interface: used to construct the function mapping relationship between each clinical parameter within each data class based on each clinical parameter of the patient under each normalized data class, and calculate the partial correlation coefficient between each clinical parameter based on the function mapping relationship;

[0091] It is worth noting that, in order to reveal the intrinsic relationship between these clinical parameters, the following multiple linear regression model was constructed:

[0092]

[0093] in, This represents the i-th clinical parameter (target variable); This represents the j-th clinical parameter (predictor variable); The intercept term represents the i-th model; This represents the influence coefficient of the j-th clinical parameter on the i-th clinical parameter; The value represents the residual term, indicating the portion not explained by the model; n represents the total number of clinical parameters involved in the modeling (6 in this case).

[0094] Based on the above, the simple correlation coefficient between any two clinical parameters is then calculated. The specific calculation formula is as follows:

[0095]

[0096] in, , These represent the sample observation values ​​of the i-th and j-th clinical parameters, respectively. , Corresponding sample mean;

[0097] Then, the partial correlation coefficients between the various clinical parameters were calculated, as follows:

[0098] The partial correlation coefficient reflects the net correlation between two variables, after controlling for the influence of other variables. Let:

[0099] In the control set The partial correlation coefficient between the i-th and j-th clinical parameters under the influence of all variables; it is worth noting that unless otherwise specified. If so, it will control the influence of all other variables by default;

[0100] The specific calculation formula is as follows:

[0101]

[0102] in, This represents the simple correlation coefficient between the i-th clinical parameter and the j-th clinical parameter; The simple correlation coefficient between the i-th clinical parameter and the k-th clinical parameter is represented by . The simple correlation coefficient between the j-th clinical parameter and the k-th clinical parameter is represented by . Represents the set of all other clinical parameters, i.e., except and All variables outside of; Representative All variables in The coefficient of determination when performing regression; Representative All variables in The coefficient of determination when performing regression;

[0103] Furthermore, the coefficient of determination represents the proportion of variance explained by the model to the total variance, and its calculation formula is as follows: in: This represents the sum of squared residuals, and the calculation process is as follows:

[0104] The total sum of squares is expressed as follows:

[0105] In the above, This represents the actual observation of the k-th sample, derived from raw clinical data, such as in... When is the target variable, it represents the CAT score of the kth patient; The model prediction value representing the k-th sample is derived from the regression equation based on other parameters of the patient. represents the average of all sample observations; where m represents the sample size.

[0106] Contribution factor acquisition interface: used to determine the direction and intensity of the influence of each clinical parameter on anesthesia risk based on the partial correlation coefficient, and to generate the contribution factor of each clinical parameter for each patient under each type of data;

[0107] It is worth noting that the sign (positive / negative) of the partial correlation coefficient represents the direction of influence; a positive value indicates a positive influence, and a negative value indicates a negative influence. Furthermore, the absolute value of the partial correlation coefficient represents the intensity of the influence. Specifically, the generation process of the contribution factor for each clinical parameter is as follows:

[0108]

[0109] in, The contributing factor representing the i-th clinical parameter; ( ) represents the sign-reset function.

[0110] Original anesthesia risk score acquisition interface: This is used to perform a product operation on each clinical parameter contribution factor of the patient under each data category and each normalized clinical parameter of the patient under each data category to obtain the original anesthesia risk score of the patient under each data category.

[0111] Anesthesia Risk Value Acquisition Interface: Used to perform non-linear transformation on the raw anesthesia risk score to obtain the anesthesia risk value of the patient under each data category;

[0112] It is worth noting that when performing nonlinear transformation, this application prioritizes using the Sigmoid function to perform nonlinear transformation, so that the output value falls within the [0,1] interval.

[0113] The anesthesia risk value matrix construction sub-interface is used to construct an anesthesia risk vector set consisting of the anesthesia risk values ​​of patients under each data category, and to form an anesthesia risk value matrix based on the anesthesia risk vector set.

[0114] It is worth noting that the formation of the anesthesia risk value matrix depends entirely on the calculation results of the anesthesia risk values ​​mentioned above. First, the following anesthesia risk vector is constructed:

[0115]

[0116] in, This represents the anesthesia risk vector for the p-th patient. represents the anesthesia risk value (range [0,1]) for the p-th patient in the m-th data category; m represents the total number of data categories.

[0117] Subsequently, the anesthesia risk vectors of all sample patients were stacked horizontally to form an anesthesia risk value matrix:

[0118] R

[0119] Where R represents the anesthesia risk value matrix.

[0120] The correlation coefficient matrix generation sub-interface is used to calculate the Pearson correlation coefficient between each type of data based on the anesthesia risk value matrix and generate the correlation coefficient matrix.

[0121] Subsequently, based on the aforementioned anesthesia risk value matrix, the Pearson correlation coefficient between any two data categories is calculated, as follows:

[0122]

[0123] in, The Pearson correlation coefficient represents the relationship between data in class i and class j. This represents the anesthesia risk value for the p-th patient under the i-th data type. This represents the anesthesia risk value for the p-th patient under the j-th data category; This represents the average anesthesia risk value for all patients in the i-th data category; represents the average anesthesia risk value for all patients in the j-th data category; n represents the sample size.

[0124] Then, all the Pearson correlation coefficients between each pair were arranged in order, resulting in the following correlation coefficient matrix:

[0125] C

[0126] Where C represents the correlation coefficient matrix.

[0127] Eigenvalue decomposition sub-interface: Used to perform eigenvalue decomposition on the correlation coefficient matrix and extract the eigenvalues ​​and corresponding eigenvectors from the correlation coefficient matrix;

[0128] Then, eigenvalue decomposition is performed on the correlation coefficient matrix: in The eigenvector corresponding to the k-th eigenvalue; Let represent the k-th eigenvalue. By solving this equation, a set of eigenvalues ​​can be obtained. , , , and its corresponding eigenvectors , , , .

[0129] Principal Component Factor Extraction Sub-interface: Used to sort eigenvalues ​​in descending order, calculate the variance contribution rate and cumulative variance contribution rate of each eigenvalue, and select the eigenvectors corresponding to the first k eigenvalues ​​whose cumulative variance contribution rate reaches a preset threshold as principal component factors;

[0130] Next, the eigenvalues ​​are sorted in descending order of magnitude, such as... ≥ ≥ , Then, the variance contribution rate of the k-th eigenvalue is calculated: And the cumulative variance contribution rate:

[0131]

[0132] in, The variance contribution rate of the kth eigenvalue; The variance contribution rate up to the k-th feature value represents the cumulative variance contribution rate; m represents the total number of data categories. The variance contribution rate of the i-th eigenvalue;

[0133] Subsequently, a cumulative variance contribution rate threshold of 85% was set, and a selection was made that... The smallest k with ≥85% is used to form the principal component factor matrix by taking the top k eigenvectors, as shown below: .

[0134] Loading coefficient extraction sub-interface: used to extract loading coefficients from principal component factors and map the loading coefficients back to the data categories that constitute the anesthesia risk value matrix, and identify the data categories that have a significant impact on the principal components;

[0135] Subsequently, loading coefficients are extracted, where each feature vector in the principal component factors corresponds to the loading coefficient of the original data category. The larger the absolute value of the loading coefficient, the more significant the influence of the data category on the corresponding principal component. For example, if the loading coefficient of a principal component on "respiratory system related data" is 0.91, it means that the principal component mainly reflects respiratory system risk; if the loading coefficient of a principal component on "exercise tolerance and metabolic function" is 0.85, it means that the principal component has a large influence on this category. In this way, the key data categories that dominate anesthesia risk can be identified.

[0136] The second weight calculation sub-interface is used to calculate the second weight of each type of data based on the variance contribution rate of the principal component factors and the corresponding loading coefficients.

[0137] Specifically, the second weight of each data class can be calculated by multiplying the variance contribution rate of the principal component factor with the corresponding loading coefficient. However, in order to ensure that the sum of the weights is 1, normalization is still required.

[0138] Furthermore, the anesthesia risk value matrix construction sub-interface to the second weight calculation sub-interface proposes a complete principal component analysis (PCA) workflow, covering multiple steps such as anesthesia risk vector set construction, correlation coefficient matrix calculation, eigenvalue decomposition, principal component factor selection, loading coefficient mapping, and weight generation. This provides a theoretical basis and technical support for variable correlation calculation in anesthesia risk assessment. The technical solution first constructs an anesthesia risk vector set and forms an anesthesia risk value matrix, integrating the multidimensional risk information of individual patients into a unified structured dataset, laying the foundation for subsequent correlation analysis. Subsequently, the system calculates the Pearson correlation coefficient between each type of data and generates a correlation coefficient matrix, revealing the linear correlation between different types of data. This helps identify whether there is redundant information or co-changing trends, providing a basis for dimensionality reduction analysis. In the PCA stage, the system performs eigenvalue decomposition on the correlation coefficient matrix to extract principal component factors. This step identifies the main factors driving changes in anesthesia risk, reducing redundant information and improving computational efficiency. Subsequently, the system selects principal components by cumulative variance contribution rate, effectively controlling computational complexity and avoiding redundant calculations while retaining maximum information, reflecting the design philosophy of "dimensionality reduction without distortion." The system further extracts loading coefficients from the principal component factors and maps them back to the original data categories, achieving a quantitative description of the role of each data category in the principal components. This helps identify which data categories contribute more to the overall risk change and has clear clinical significance. Finally, the system calculates the second weight based on the variance contribution rate and loading coefficients. This method comprehensively considers the importance of the principal components and their influence on the original data categories, resulting in weights with high scientific validity and robustness. Therefore, the technical effects brought about by the sub-interface for constructing the anesthesia risk value matrix and the sub-interface for calculating the second weight include enhanced interpretability of the score, improved stability of the calculation, support for personalized weight generation, and theoretical support for the scoring system. It is an indispensable technical support for achieving intelligent, accurate, and personalized anesthesia risk assessment, providing the entire system with a rigorous, logically clear, and mathematically sound correlation calculation and weight generation mechanism. It is an important component of this invention for achieving scientific and efficient anesthesia risk assessment.

[0139] The comprehensive anesthesia risk value calculation subunit is used to combine the final weight of each type of data with the anesthesia risk value of the patient under each type of data to obtain the patient's comprehensive anesthesia risk value.

[0140] The Comprehensive Anesthesia Risk Score Acquisition Subunit is used to normalize the patient's comprehensive anesthesia risk value to obtain the patient's comprehensive anesthesia risk score.

[0141] Optimized score acquisition subunit: This unit combines the patient's comprehensive anesthesia risk score with the patient's initial anesthesia risk score and takes the average to obtain the final anesthesia risk score for the patient.

[0142] In this invention, the clinical parameter extraction interface and the anesthesia risk value acquisition interface undertake the most critical risk value calculation task in the anesthesia risk scoring system. A complete calculation process is proposed to accurately assess the patient's anesthesia risk level for each data type. This process includes steps such as clinical parameter extraction, normalization processing, function mapping relationship construction, partial correlation coefficient analysis, contribution factor generation, product operation, and nonlinear transformation, providing a solid mathematical foundation and operational support for the accuracy of the scoring system. First, the system ensures the completeness and representativeness of the input parameters by extracting clinical parameters for each data type, providing the original basis for subsequent risk value calculation. Then, normalization processing solves the calculation bias problem caused by inconsistencies in the dimensions and value ranges of different parameters, ensuring all parameters are on a uniform scale and improving the stability and generalization ability of the calculation process. In the function mapping stage, the system constructs mapping logic based on the interrelationships between parameters and eliminates interference from other parameters through partial correlation coefficient analysis, thereby obtaining the true correlation between two parameters and providing a basis for the subsequent generation of contribution factors. Based on the influence direction and intensity of the partial correlation coefficient, the system generates a contribution factor for each parameter, clarifying its specific influence on anesthesia risk. Based on this, the system performs a term-by-term multiplication of the contribution factors with the normalized parameters to obtain the original anesthesia risk score, simulating the cumulative effect of parameters on risk. Finally, a nonlinear transformation is used to standardize the original score, mapping it to the [0,1] interval to form the final anesthesia risk value. This nonlinear processing method not only enhances the comparability of the output but also better matches the clinical need for "low, medium, and high" risk level classification. Therefore, the clinical parameter extraction interface to the anesthesia risk value acquisition interface provides a data-driven, statistically based, and nonlinear mapping-output method for calculating anesthesia risk values. Its advantages lie in its ability to quantitatively assess the actual impact of each parameter, control interference factors between parameters, achieve standardized output of risk values, and improve the interpretability and clinical applicability of the scores. It is one of the most critical technical cores of the entire visit assessment system.

[0143] Anesthesia planning module: Connected to the data analysis module, it is used to develop anesthesia plans based on the final patient's anesthesia risk score;

[0144] It is worth noting that after completing the preoperative visit and assessment, the final anesthesia risk score range of the patient output by the system is [0,1], which reflects the overall risk level of complications or adverse reactions that the patient may experience during anesthesia. This embodiment aims to illustrate how to develop a targeted, safe and individualized anesthesia plan based on this score, thereby improving the safety of the anesthesia process and the scientific nature of intraoperative management.

[0145] Typically, the system will classify patients into three risk levels based on their final anesthesia risk score, as detailed in the risk level table below:

[0146]

[0147] This risk classification is based on clinical expert consensus and historical data statistical analysis results to ensure a high degree of match with actual clinical risks. Subsequently, the system automatically generates the following core anesthesia plan content according to the patient's risk level, including: anesthesia method selection, anesthetic drug selection and dosage control, preoperative preparation suggestions, key points of intraoperative monitoring, and postoperative recovery management suggestions.

[0148] It should be noted that the terminology used in this invention is for describing specific embodiments only and is not intended to limit the scope of this application. As shown in this specification, unless the context clearly indicates otherwise, words such as "a," "an," "an," and / or "the" do not specifically refer to the singular and may include the plural. The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, 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, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, or apparatus that includes said element.

[0149] It should also be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Unless otherwise expressly specified and limited, the terms "installed," "connected," "linked," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components. For those skilled in the art, the specific meaning of the above terms in the present invention can be understood according to the specific circumstances.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A preoperative visit assessment system for anesthesia risk, characterized by, Comprise the following modules: Data acquisition module: for patients to scan the code through the two-dimensional code to log in the system and fill in the personal basic information data; call the personal basic information data, generate the body robustness index; if the current patient's body strength index is less than the decision threshold, then execute the subsequent score optimization link; if the current patient's body strength index is greater than or equal to the decision threshold, then skip the score optimization, directly use the initial score; Data analysis module: connected with the data acquisition module, the data analysis module comprises a maximum weight determination subunit for analyzing the patient's personal basic information data according to the maximum weight determination subunit to obtain the final patient's anesthesia risk score, comprising: Index calculation unit: for calculating the patient's chronic obstructive pulmonary disease CAT index, coagulation disorder tendency index, MET activity equivalent index, and dyspnea index based on the patient's personal basic information data; Initial anesthesia risk score calculation unit: for synthesizing the chronic obstructive pulmonary disease CAT index, coagulation disorder tendency index, MET activity equivalent index, and dyspnea index to obtain the patient's initial anesthesia risk score; Score optimization unit: for optimizing the patient's initial anesthesia risk score to obtain the final patient's anesthesia risk score, comprising: Data classification subunit: for classifying the patient's personal basic information data according to data types; Risk value calculation subunit: for calculating the patient's anesthesia risk value under each type of data; Comprehensive anesthesia risk value calculation subunit: for combining the maximum weight of each type of data with the patient's anesthesia risk value under each type of data to obtain the patient's comprehensive anesthesia risk value; Comprehensive anesthesia risk score acquisition subunit: for normalizing the patient's comprehensive anesthesia risk value to obtain the patient's comprehensive anesthesia risk score; Optimized score acquisition subunit: for combining the patient's comprehensive anesthesia risk score with the patient's initial anesthesia risk score and taking the average to obtain the final patient's anesthesia risk score; Maximum weight determination subunit: for determining the maximum weight of each type of data based on the patient's anesthesia risk value under each type of data, the maximum weight determination subunit comprises the following interfaces; First weight calculation interface: for determining the first weight of each type of data according to the ratio of the patient's anesthesia risk value under each type of data to the sum of the patient's anesthesia risk values under all types of data; Second weight determination interface: for analyzing the correlation between the patient's anesthesia risk values under each type of data and determining the second weight of each type of data based on the analysis result, comprising: Anesthesia risk value matrix construction sub-interface: for constructing an anesthesia risk vector set composed of the patient's anesthesia risk values under each type of data, and forming an anesthesia risk value matrix based on the anesthesia risk vector set; Correlation coefficient matrix generation sub-interface: for calculating the Pearson correlation coefficient between each type of data based on the anesthesia risk value matrix and generating a correlation coefficient matrix; Eigenvalue decomposition sub-interface: for eigenvalue decomposition of the correlation coefficient matrix and extracting the eigenvalues in the correlation coefficient matrix and the eigenvectors corresponding to the eigenvalues; The principal component factor extraction sub-interface is configured to sort the eigenvalues in descending order, calculate the variance contribution rate and the cumulative variance contribution rate of each eigenvalue, select the eigenvectors corresponding to the first k eigenvalues whose cumulative variance contribution rate reaches a preset threshold as the principal component factors, and the like. The load coefficient extraction sub-interface is configured to extract the load coefficients from the principal component factors, map the load coefficients back to the data categories constituting the anesthesia risk value matrix, and identify the data categories that have a significant impact on the principal components. The second weight calculation sub-interface is configured to calculate the second weight of each data category according to the variance contribution rate of the principal component factors and the corresponding load coefficients. The maximum weight calculation interface is configured to take the mean of the sum of the first weight of each data category and the second weight of each data category as the maximum weight of each data category. The anesthesia plan making module is connected with the data analysis module and is configured to make an anesthesia plan based on the anesthesia risk score of the final patient.

2. The preoperative anesthetic risk visit assessment system of claim 1, wherein, The calculation of the anesthesia risk value of the patient under each data category includes: The clinical parameter extraction interface is configured to extract the clinical parameters of the patient under each data category. The clinical parameter normalization interface is configured to perform normalization processing on each clinical parameter of the patient under each data category. The partial correlation coefficient calculation interface is configured to construct a functional mapping relationship between the clinical parameters in each data category based on each clinical parameter of the patient under each data category after normalization, and calculate the partial correlation coefficient between the clinical parameters based on the functional mapping relationship. The contribution factor acquisition interface is configured to determine the influence direction and intensity of each clinical parameter on the anesthesia risk according to the partial correlation coefficient, and generate the contribution factor of each clinical parameter of the patient under each data category. The original anesthesia risk score acquisition interface is configured to perform item-by-item multiplication operation on the contribution factor of each clinical parameter of the patient under each data category and each clinical parameter of the patient under each data category after normalization, to obtain the original anesthesia risk score of the patient under each data category. The anesthesia risk value acquisition interface is configured to perform nonlinear transformation on the original anesthesia risk score to obtain the anesthesia risk value of the patient under each data category.

Images