Anesthesia risk assessment and early warning system for perioperative patients based on big data analysis
Through big data analysis and machine learning technology, a big data-based anesthesia risk assessment system has been established, which solves the problem of inaccurate anesthesia risk assessment in the existing technology, achieves more accurate risk prediction and surgical plan optimization, and reduces surgical risks .
Patent Information
- Application Number
- CN202311166252.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-11
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2043-09-11
AI Technical Summary
Existing anesthesia risk assessment methods such as ASA grading standards are not accurate enough to predict the risk of anesthesia surgery, lack intelligence and personalization, and it is difficult to effectively reduce the potential risks during the perioperative period.
Using anesthesia risk assessment and early warning system based on big data analysis, the patient's basic physical indicators, laboratory indicators and surgical information are optimized through data preprocessing and machine learning models, and the ratio Q and predictors of categorical variables are calculated using formulas to establish logic The regression model conducts risk prediction, and visualizes and warnings of abnormal results.
A more accurate prediction of risk for anesthesia surgery has been achieved, helping doctors formulate and optimize surgical plans, reduce surgical risks, and improve surgical safety and medical quality.
Smart Images
Figure CN117198533B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of perioperative anesthesia risk assessment, and specifically relates to a perioperative patient anesthesia risk assessment and early warning system based on big data analysis. Background Art
[0002] With the development of modern medicine, the number of surgeries in hospitals has increased significantly, and more and more hospitals have paid special attention to the safety in the operating room. As anesthesia is a key link in the surgical process, the risk assessment and safety management of anesthesia have become the focus of medical treatment today. Although anesthesia medicine has been developed for more than a century, various adverse events of anesthesia still occur from time to time. The main reason is that the use of anesthetic drugs is relatively complex, and the medical process of anesthesia itself also has certain risk factors. At present, preoperative risk assessment of patients and early warning of disease complications and anesthesia risks can effectively prevent and avoid risks, which is an important means to ensure anesthesia safety.
[0003] The ASA anesthesia risk classification standard is currently the most commonly used classification method. This classification standard classifies patients according to their physical condition and surgical risk before anesthesia, and divides patients into six levels to guide the risks of anesthesia surgery in practice. However, this classification standard is relatively rough in assessing the patient's condition and is not accurate enough in predicting the risks of anesthesia surgery. The industry still needs a smarter method for anesthesia surgery to provide effective predictions, modify the anesthesia surgery plan based on risk rating, take preventive measures, reduce potential perioperative risks, and improve surgical safety and medical quality. Summary of the invention
[0004] The purpose of the present invention is to solve the problems existing in the above-mentioned prior art and to provide a perioperative patient anesthesia risk assessment and early warning system based on big data analysis, which can accurately assess and predict the risks of anesthesia surgery and reduce potential risks of surgery.
[0005] The present invention is achieved through the following technical solutions:
[0006] Embodiment 1 provides a perioperative patient anesthesia risk assessment and early warning system based on big data analysis, including:
[0007] 1) Data acquisition unit: acquires the original data of several patients from a medical database as the initial data set;
[0008] 2) Data processing unit: performing data preprocessing on the initial data set to obtain a basic data set, wherein the data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables; wherein, for nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1):
[0009] Q = log [ P ( X i = x i | E = 1 ) / P ( X i = x i | E = 0 ) ] (1)
[0010] and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2),
[0011] (2)
[0012] Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results;
[0013] 3) Model prediction unit: trains and optimizes the machine learning model using the basic data set, and the model is defined as:
[0014] ln [ P A P B ] = β 0 , 1 + β 1 , 1 X ( 1 ) + β 2 , 1 X ( 2 ) + ... + β m , 1 X ( m ) = y 1 (3)
[0015] ln [ P B P C ] = β 0 , 2 + β 1 , 2 X ( 1 ) + β 2 , 2 X ( 2 ) + ... + β m , 2 X ( m ) = y 2 (4)
[0016] Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor; P A = P [ j = A | X ] , P B = P [ j = B | X ] , P C = P [ j = C | X ] , ; , Respectively represent ln [ P A P B ] and ln [ P B P C ] Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6):
[0017] (5)
[0018] (6)
[0019] 4) Result display unit: Visually display the results output by the model prediction unit, predict the patient's perioperative anesthesia risk probability, and issue warnings for abnormal results.
[0020] Furthermore, the original data includes the patient's basic physical indicators, laboratory test indicators, surgical information, and ASA classification given by the doctor;
[0021] Furthermore, the basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature;
[0022] Furthermore, the laboratory test indicators include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar;
[0023] Furthermore, the surgical information includes surgical type, surgical grade, organ status, and blood loss.
[0024] Furthermore, the removal of outliers includes removing outliers that are considered unreasonable by medical experts, and removing observations located at the top and bottom 1% in the distribution of continuous variables;
[0025] Furthermore, the estimating missing data includes estimating missing values using an automatic algorithm.
[0026] Furthermore, the visual display includes displaying the risk probability on a display screen and recording the risk probability in an initial data set of a medical database.
[0027] Embodiment 2 provides a perioperative patient anesthesia risk assessment and early warning method based on big data analysis, including:
[0028] 1) Obtaining raw data: Obtaining raw data of several patients from a medical database as the initial data set;
[0029] 2) Raw data processing: The initial data set is preprocessed to obtain a basic data set. The data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables. For nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1):
[0030] Q = log [ P ( X i = x i | E = 1 ) / P ( X i = x i | E = 0 ) ] (1)
[0031] and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2),
[0032] (2)
[0033] Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results;
[0034] 3) Model training: The basic data set is used to train and optimize the machine learning model, and the model is defined as:
[0035] ln [ P A P B ] = β 0 , 1 + β 1 , 1 X ( 1 ) + β 2 , 1 X ( 2 ) + ... + β m , 1 X ( m ) = y 1 (3)
[0036] ln [ P B P C ] = β 0 , 2 + β 1 , 2 X ( 1 ) + β 2 , 2 X ( 2 ) + ... + β m , 2 X ( m ) = y 2 (4)
[0037] Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor; P A = P [ j = A | X ] , P B = P [ j = B | X ] , P C = P [ j = C | X ] , ; , Respectively represent ln [ P A P B ] and ln [ P B P C ] Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6):
[0038] (5)
[0039] (6)
[0040] 4) Visual display: Visualize the output of the model to predict the patient's perioperative anesthesia risk probability and issue warnings for abnormal results.
[0041] Preferably, the original data includes the patient's basic physical indicators, laboratory test indicators, surgical information, and ASA classification given by the doctor;
[0042] Preferably, the basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature;
[0043] Preferably, the laboratory test indicators include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar;
[0044] Preferably, the surgical information includes surgical type, surgical grade, organ status, and blood loss.
[0045] Preferably, the removing of abnormal values includes removing outliers that are considered unreasonable by medical experts, and removing the observations located at the top and bottom 1% in the distribution of continuous variables;
[0046] Preferably, said estimating missing data comprises estimating missing values using an automatic algorithm.
[0047] Preferably, the visual display includes displaying the risk probability on a display screen and simultaneously recording the risk probability in an initial data set of a medical database.
[0048] Compared with the existing technology, the present invention provides a perioperative patient anesthesia risk assessment and early warning system and method based on big data analysis. By using machine learning technology, it can realize personalized and targeted anesthesia risk assessment, more accurately predict the patient's anesthesia surgery risk, help doctors formulate and optimize surgical plans, and reduce surgical risks. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 This is a schematic diagram of an anesthesia risk assessment and early warning system for perioperative patients based on big data analysis.
[0050] Figure 2 This is a schematic diagram of an anesthesia risk assessment and early warning method for perioperative patients based on big data analysis. DETAILED DESCRIPTION
[0051] The present invention is further described in detail below in conjunction with the accompanying drawings:
[0052] Embodiment 1
[0053] like Figure 1 As shown, a perioperative patient anesthesia risk assessment and early warning system based on big data analysis is provided, including:
[0054] 1) Data acquisition unit: acquires the original data of several patients from a medical database as the initial data set;
[0055] 2) Data processing unit: performing data preprocessing on the initial data set to obtain a basic data set, wherein the data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables; wherein, for nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1):
[0056] Q = log [ P ( X i = x i | E = 1 ) / P ( X i = x i | E = 0 ) ] (1)
[0057] and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2),
[0058] (2)
[0059] Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results;
[0060] 3) Model prediction unit: trains and optimizes the machine learning model using the basic data set, and the model is defined as:
[0061] ln [ P A P B ] = β 0 , 1 + β 1 , 1 X ( 1 ) + β 2 , 1 X ( 2 ) + ... + β m , 1 X ( m ) = y 1 (3)
[0062] ln [ P B P C ] = β 0 , 2 + β 1 , 2 X ( 1 ) + β 2 , 2 X ( 2 ) + ... + β m , 2 X ( m ) = y 2 (4)
[0063] Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor; P A = P [ j = A | X ] , P B = P [ j = B | X ] , P C = P [ j = C | X ] , ; , Respectively represent ln [ P A P B ] and ln [ P B P C ] Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6):
[0064] (5)
[0065] (6)
[0066] 4) Result display unit: Visually display the results output by the model prediction unit, predict the patient's perioperative anesthesia risk probability, and issue warnings for abnormal results.
[0067] Among them, the original data includes the patient's basic physical indicators, laboratory indicators, surgical information and ASA grade given by the doctor; the basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature; the laboratory indicators include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar; the surgical information includes surgical type, surgical grade, organ status, and blood loss.
[0068] Removing outliers includes eliminating outliers that are considered unreasonable by medical experts, as well as eliminating the top and bottom 1% of observations in the distribution of continuous variables; imputing missing data includes using automatic algorithms to impute missing values.
[0069] The visual display includes displaying the risk probability on a display screen and recording the risk probability in an initial data set of a medical database.
[0070] Embodiment 2
[0071] like Figure 2 As shown, a perioperative patient anesthesia risk assessment and early warning method based on big data analysis is provided, including:
[0072] 1) Obtaining raw data: Obtaining raw data of several patients from a medical database as the initial data set;
[0073] 2) Raw data processing: The initial data set is preprocessed to obtain a basic data set. The data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables. For nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1):
[0074] Q = log [ P ( X i = x i | E = 1 ) / P ( X i = x i | E = 0 ) ] (1)
[0075] and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2),
[0076] (2)
[0077] Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results;
[0078] 3) Model training: The basic data set is used to train and optimize the machine learning model, and the model is defined as:
[0079] ln [ P A P B ] = β 0 , 1 + β 1 , 1 X ( 1 ) + β 2 , 1 X ( 2 ) + ... + β m , 1 X ( m ) = y 1 (3)
[0080] ln [ P B P C ] = β 0 , 2 + β 1 , 2 X ( 1 ) + β 2 , 2 X ( 2 ) + ... + β m , 2 X ( m ) = y 2 (4)
[0081] Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor; P A = P [ j = A | X ] , P B = P [ j = B | X ] , P C = P [ j = C | X ] , ; , Respectively represent ln [ P A P B ] and ln [ P B P C ] Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6):
[0082] (5)
[0083] (6)
[0084] 4) Visual display: Visualize the output of the model to predict the patient's perioperative anesthesia risk probability and issue warnings for abnormal results.
[0085] Among them, the original data includes the patient's basic physical indicators, laboratory indicators, surgical information and ASA grade given by the doctor; the basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature; the laboratory indicators include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar; the surgical information includes surgical type, surgical grade, organ status, and blood loss.
[0086] Removing outliers includes eliminating outliers that are considered unreasonable by medical experts, as well as eliminating the top and bottom 1% of observations in the distribution of continuous variables; imputing missing data includes using automatic algorithms to impute missing values.
[0087] The visual display includes displaying the risk probability on a display screen and recording the risk probability in an initial data set of a medical database.
[0088] Compared with the existing technology, the present invention provides a perioperative patient anesthesia risk assessment and early warning system and method based on big data analysis. By using machine learning technology, it can realize personalized and targeted anesthesia risk assessment, more accurately predict the patient's anesthesia surgery risk, help doctors formulate and optimize surgical plans, and reduce surgical risks.
[0089] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0090] In the description of the present invention, unless otherwise specified, the terms "upper", "lower", "left", "right", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They 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 direction, be constructed and operated in a specific direction. Therefore, they cannot be understood as limitations on the present invention.
[0091] Finally, it should be explained that the above technical solution is only one implementation method of the present invention. For those skilled in the art, it is easy to make various types of improvements or modifications based on the application methods and principles disclosed in the present invention, and it is not limited to the method described in the above specific implementation method of the present invention. Therefore, the method described above is only preferred and does not have a restrictive meaning.
Claims
1. A perioperative anesthesia risk assessment and early warning system based on big data analysis, including: 1) Data acquisition unit: acquires the original data of several patients from a medical database as the initial data set; The original data includes the patient's basic physical indicators, laboratory test indicators, surgical information, and the ASA grade given by the doctor; 2) Data processing unit: performing data preprocessing on the initial data set to obtain a basic data set, wherein the data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables; wherein, for nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1): （1）； and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2), （2）； Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results; 3) Model prediction unit: trains and optimizes the machine learning model using the basic data set, and the model is defined as: （3）； （4）； Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor, let , , , ; , Respectively represent and Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6): （5）； （6）； 4) Result display unit: Visually display the results output by the model prediction unit, predict the patient's perioperative anesthesia risk probability, and issue warnings for abnormal results.
2. According to claim 1, a perioperative patient anesthesia risk assessment and early warning system based on big data analysis is characterized by: The basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature; The laboratory parameters include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar; The surgical information includes surgical type, surgical grade, organ status, and blood loss.
3. The perioperative patient anesthesia risk assessment and early warning system based on big data analysis according to claim 1 is characterized by: The removal of outliers includes removing outliers that are considered unreasonable by medical experts, and removing observations at the top and bottom 1% in the distribution of continuous variables; The estimating missing data includes estimating missing values using an automatic algorithm.
4. The perioperative patient anesthesia risk assessment and early warning system based on big data analysis according to claim 1 is characterized by: The visual display includes displaying the risk probability on a display screen and recording the risk probability in an initial data set of a medical database.
5. A perioperative patient anesthesia risk assessment and early warning method based on big data analysis, including: 1) Obtaining raw data: Obtaining raw data of several patients from a medical database as the initial data set; The original data includes the patient's basic physical indicators, laboratory test indicators, surgical information, and the ASA grade given by the doctor; 2) Raw data processing: The initial data set is preprocessed to obtain a basic data set. The data preprocessing includes removing outliers, estimating missing data, and optimizing categorical variables and nominal variables. For nominal variables and categorical variables with more than two levels, each level is replaced by the ratio Q calculated by formula (1): （1）； and Representing positive and negative results respectively. represents a categorical variable, is the estimated value using formula (2), （2）； Represents categorical variables The results at level j, Indicates appearance and The number of times; The value is 0 or 1, indicating positive and negative results; 3) Model training: The basic data set is used to train and optimize the machine learning model, and the model is defined as: （3）； （4）； Here, set the categorical variable With levels A, B and C, m predictors X (1) , X (2) , …, X (m) , , , …, , , , …, is the influencing factor of each predictor; , , , ; , Respectively represent and Based on the above formulas (3) and (4), all possibilities are calculated by the following formulas (5) and (6): （5）； （6）； 4) Visual display: Visualize the output of the model to predict the patient's perioperative anesthesia risk probability and issue warnings for abnormal results.
6. The method for perioperative patient anesthesia risk assessment and early warning based on big data analysis according to claim 5, characterized in that: The basic physical indicators include age, weight, BMI, heart rate, blood pressure, respiration, and body temperature; The laboratory parameters include sodium, potassium, creatinine, urea nitrogen, hemoglobin, hematocrit, and blood sugar; The surgical information includes surgical type, surgical grade, organ status, and blood loss.
7. The method for perioperative patient anesthesia risk assessment and early warning based on big data analysis according to claim 5, characterized in that: The removal of outliers includes removing outliers that are considered unreasonable by medical experts, and removing observations at the top and bottom 1% in the distribution of continuous variables; The estimating missing data includes estimating missing values using an automatic algorithm.
8. The method for perioperative anesthesia risk assessment and early warning based on big data analysis according to claim 5, characterized in that: The visual display includes displaying the risk probability on a display screen and recording the risk probability in an initial data set of a medical database.