Lung infection risk probability prediction method, device, equipment, medium and product

Abstract

The invention discloses a pulmonary infection risk probability prediction method and device, equipment, a medium and a product, and relates to the field of pulmonary infection risk probability prediction.The method comprises the steps that a training data set is acquired; based on the training data set, judging whether a pulmonary infection risk prediction probability value of a pulmonary infection risk probability prediction model and an actual risk probability value are in a set error range or not; if yes, predicting a pulmonary infection risk probability value of the postoperative patient according to the pulmonary infection risk probability prediction model; if not, correcting the pulmonary infection risk probability prediction model; and predicting the pulmonary infection risk probability value of the postoperative patient according to the corrected pulmonary infection risk probability prediction model. According to the application, the accuracy of predicting the pulmonary infection risk probability is improved.

Description

Technical Field

[0001] The present application relates to the field of lung infection risk probability prediction, and in particular to a lung infection risk probability prediction method, device, equipment, medium and product. Background Art

[0002] Hip fractures mainly refer to fractures that occur within 5 cm from the plane of the femoral head to the lesser trochanter of the femur. Clinically, they can be divided into femoral neck fractures, intertrochanteric fractures and subtrochanteric fractures. They often occur in the elderly. As a major public health issue, hip fractures have received much attention from countries around the world. Due to the accelerated aging of the population and the increase in the elderly population, the incidence of hip fractures has increased year by year. The incidence of hip fractures in the elderly over 60 years old exceeds 214.89 / 100,000.

[0003] The main cause of death after surgery for elderly people with hip fractures is lung infection. Timely prediction of the risk of lung infection in patients and effective medical and nursing interventions are the key to preventing lung infection and promoting patient recovery.

[0004] There have been studies on prediction models for postoperative lung infection in elderly patients with hip fractures. However, the predicted probability values ​​of lung infection risk of existing lung infection prediction models have large errors compared with the actual risk probability values, and they are unable to accurately identify people at high risk of lung infection after hip fracture surgery. Summary of the invention

[0005] The purpose of this application is to provide a method, device, equipment, medium and product for predicting the probability of lung infection risk, so as to solve the problem of low accuracy of the existing lung infection prediction model in predicting the probability of lung infection risk.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a method for predicting the probability of lung infection risk, comprising:

[0008] Obtaining a training data set; the training data set includes multiple groups of data, each group of data includes age, gender, white blood cell count, albumin-globulin ratio, hemoglobin, and actual risk probability value of lung infection;

[0009] Based on the training data set, determining whether the predicted probability value of the lung infection risk of the lung infection risk probability prediction model and the actual risk probability value are within a set error range; the lung infection risk probability prediction model is established with the age, the gender, the white blood cell count, the albumin-globulin ratio and the hemoglobin as inputs and the actual lung infection risk probability value as output;

[0010] If yes, predicting the lung infection risk probability value of the postoperative patient according to the lung infection risk probability prediction model;

[0011] If not, calibrating the lung infection risk probability prediction model;

[0012] The lung infection risk probability value of postoperative patients was predicted based on the corrected lung infection risk probability prediction model.

[0013] Optionally, the lung infection risk probability prediction model is corrected by using an intercept correction method, including:

[0014] Based on the training data set, a correction factor is calculated according to the predicted probability value of the lung infection risk of the lung infection risk probability prediction model and the actual probability value of the lung infection risk;

[0015] Determining a corrected intercept according to the correction factor and the intercept of the lung infection risk probability prediction model;

[0016] The lung infection risk probability prediction model is corrected according to the corrected intercept, and a corrected lung infection risk probability prediction model is established.

[0017] Optionally, the correction factor = Ln((actual risk probability value of lung infection / (1-actual risk probability value of lung infection)) / (predicted risk probability value of lung infection / (1-predicted risk probability value of lung infection))).

[0018] Optionally, the corrected intercept α1=α0+correction factor; wherein α0 is the intercept of the lung infection risk probability prediction model.

[0019] Optionally, the corrected lung infection risk probability prediction model is: Model 1 = 0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB+α1;

[0020] Among them, Model 1 is the adjusted lung infection risk probability prediction model, which is a prediction model corrected by intercept correction; WBC is white blood cell count; AGR is albumin-globulin ratio; HGB is hemoglobin.

[0021] Optionally, the lung infection risk probability prediction model is corrected by simultaneously correcting the intercept and the regression coefficient, including:

[0022] Updating the intercept of the lung infection risk probability prediction model to generate an updated intercept;

[0023] Based on the intercept of the lung infection risk probability prediction model, the regression coefficients of all predictors are updated as a whole; the predictors include age, gender, white blood cell count, albumin-globulin ratio and hemoglobin;

[0024] The lung infection risk probability prediction model is corrected according to the updated intercept and the overall updated regression coefficient to establish a corrected lung infection risk probability prediction model.

[0025] Optionally, the corrected lung infection risk probability prediction model is: Model 2 = α2 + β*(0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB);

[0026] Among them, Model 1 is the corrected lung infection risk probability prediction model, which is a prediction model calibrated by simultaneously correcting the intercept and regression coefficient; α2 is the updated intercept; β is the overall updated regression coefficient; WBC is the white blood cell count; AGR is the albumin-globulin ratio; HGB is hemoglobin.

[0027] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any one of the above-mentioned methods for predicting the probability of lung infection risk.

[0028] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above-mentioned methods for predicting the probability of lung infection risk.

[0029] In a fourth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned methods for predicting the probability of lung infection risk.

[0030] According to the specific embodiments provided by the present application, the present application discloses the following technical effects: based on a training data set, the present application determines whether the lung infection risk prediction probability value of the lung infection risk probability prediction model and the actual risk probability value are within a set error range, thereby determining whether the lung infection risk prediction probability value predicted by the lung infection risk probability prediction model is accurate, and determining whether to calibrate the lung infection risk probability prediction model based on the judgment result, and using the corrected lung infection risk probability prediction model to predict the lung infection risk probability value of postoperative patients, thereby improving the accuracy of predicting the lung infection risk probability. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 This is a flow chart of a method for predicting the probability of lung infection risk in one embodiment of the present application;

[0033] Figure 2 This is a preview of the Model 1 online web page provided in one embodiment of the present application;

[0034] Figure 3 This is a preview of the Model 2 online web page provided in one embodiment of the present application. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0036] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0037] The method for predicting the probability of lung infection risk provided in the embodiment of the present application includes:

[0038] Step 101: Obtain a training data set; the training data set includes multiple groups of data, each group of data includes age, gender, white blood cell count, albumin-globulin ratio, hemoglobin and actual risk probability value of lung infection.

[0039] Step 102: Based on the training data set, determine whether the predicted probability value of the lung infection risk of the lung infection risk probability prediction model and the actual risk probability value are within a set error range. If so, execute step 103; if not, execute step 104; the lung infection risk probability prediction model is established with the age, the gender, the white blood cell count, the albumin-globulin ratio and the hemoglobin as inputs and the actual risk probability value of the lung infection as output.

[0040] Step 103: predicting the lung infection risk probability value of the postoperative patient according to the lung infection risk probability prediction model.

[0041] Step 104: Correcting the lung infection risk probability prediction model.

[0042] In an exemplary embodiment, the lung infection risk probability prediction model is corrected by using an intercept correction method, including: based on the training data set, calculating a correction factor according to the lung infection risk prediction probability value of the lung infection risk probability prediction model and the actual lung infection risk probability value; determining a corrected intercept according to the correction factor and the intercept of the lung infection risk probability prediction model; correcting the lung infection risk probability prediction model according to the corrected intercept to establish a corrected lung infection risk probability prediction model.

[0043] Furthermore, the correction factor = Ln((actual risk probability value of lung infection / (1-actual risk probability value of lung infection)) / (predicted risk probability value of lung infection / (1-predicted risk probability value of lung infection))).

[0044] Furthermore, the corrected intercept α1=α0+correction factor; wherein α0 is the intercept of the lung infection risk probability prediction model, specifically, α1=-2.538.

[0045] Furthermore, the adjusted lung infection risk probability prediction model is: Model 1 = 0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB+α1; wherein, Model 1 is the adjusted lung infection risk probability prediction model, and the adjusted lung infection risk probability prediction model is a prediction model corrected by intercept correction; WBC is white blood cell count; AGR is albumin-globulin ratio; HGB is hemoglobin.

[0046] In another exemplary embodiment, the lung infection risk probability prediction model is corrected by simultaneously correcting the intercept and the regression coefficient, including:

[0047] Updating the intercept of the lung infection risk probability prediction model to generate an updated intercept;

[0048] Based on the intercept of the lung infection risk probability prediction model, the regression coefficients of all predictors are updated as a whole; the predictors include age, gender, white blood cell count, albumin-globulin ratio and hemoglobin;

[0049] The lung infection risk probability prediction model is corrected according to the updated intercept and the overall updated regression coefficient to establish a corrected lung infection risk probability prediction model.

[0050] In an exemplary embodiment, the corrected lung infection risk probability prediction model is: Model2=α2+β*(0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB);

[0051] Among them, Model 1 is the corrected lung infection risk probability prediction model, which is a prediction model calibrated by simultaneously correcting the intercept and regression coefficient; α2 is the updated intercept, specifically, α2=-2.766; β is the overall updated regression coefficient; WBC is the white blood cell count; AGR is the albumin-globulin ratio; HGB is hemoglobin.

[0052] Step 105: Predicting the lung infection risk probability value of the postoperative patient according to the corrected lung infection risk probability prediction model.

[0053] Furthermore, the two corrected lung infection risk probability prediction models were compared to predict the lung infection risk probability values ​​of postoperative patients, and the larger lung infection risk probability value was used as the final lung infection risk probability value, which would help doctors formulate effective medical and nursing interventions to prevent lung infection and promote patient recovery.

[0054] The lung infection risk probability prediction method of the present application is verified by the following examples.

[0055] A total of 1341 elderly patients with hip fractures admitted to the Orthopedics Department of Shenzhen Second People's Hospital from January 2012 to December 2020 were retrospectively collected to explore the independent risk factors for postoperative pulmonary infection in patients with hip fractures and to establish a pulmonary infection risk probability prediction model. In this embodiment, the pulmonary infection risk probability prediction model is a Logistic regression prediction model. The main method is to randomly divide the patients into 939 training data sets and 402 test data sets (with a random seed of 1122) in a ratio of 7:3; after grouping, the number of cases in the pulmonary infection group in the training data set is 58, and the number of cases in the pulmonary infection group in the test data set is 21.

[0056] In the research of prediction models, the purpose of using the ROSE (Random Over-Sampling Examples) algorithm is to address the problem of class imbalance. Specifically, the ROSE algorithm is used to handle the class imbalance situation in the training dataset, where the number of samples in one class is far less than that in another class. The ROSE algorithm balances the sample distribution between different classes by synthesizing new minority class samples, thereby improving the model's recognition ability for minority class samples and enhancing the model's generalization ability and prediction performance. The application of the ROSE algorithm can help improve the model's performance on imbalanced datasets and enhance the prediction accuracy of the model.

[0057] In this application, for the pulmonary infection group and the non-postoperative pulmonary infection group, due to the relatively serious problem of sample proportion imbalance between the two groups of data, it is easy to lead to insufficient characterization ability of the model for sparse samples and a significant decline in the classification ability of sparse samples, resulting in bias in the results of the prediction model. Usually, to address the problem of binary classification imbalance data, data balance can be achieved through methods such as data sampling or data synthesis. The research results show that the proportion of patients with postoperative pulmonary infection is 5.89%, and the proportion of patients without postoperative pulmonary infection is 94.1%. There is an obvious imbalance between the two parts of patients. If a model is constructed using imbalanced data, it will affect the prediction accuracy of the model for positive cases and increase the prediction error because the model constructed using imbalanced data inadequately characterizes the minority class of positive cases. To solve the problem of imbalanced data, the random over-sampling technique, namely the ROSE algorithm, is introduced to resample the original data to obtain new sample data with the positive and negative case proportions close to 1.

[0058] In this application, data balance is mainly achieved by oversampling and synthesizing the minority class samples to increase the number of minority class samples.

[0059] The specific implementation steps are as follows: First, consider a dataset T with a sample size of N N , where yi is the label value of each sample in the data, and i = 1, 2,..., N. For binary classification problems, yi ∈ {Y0, Y1}, and let Nj < N be the number of samples in class Yj, j = 0, 1. xi ∈ R d is the vector composed of the feature corresponding data related to yi, and its probability density function f(x) position, and d is the number of features. The specific steps to generate a new data based on the ROSE algorithm are: (1) Draw y* = Yj from the data with probability πj; (2) Randomly draw data (xi, yi) from the dataset TN with probability 1 / N; (3) K Hj (., xi) to select the sample x*, K HjThe probability distribution of (., xi) is centered on xi, and the covariance matrix is ​​Hj. Based on the above (1)-(3), a new sample of category Yj based on the existing data is obtained. After repeating steps (1)-(3) M times, a new synthetic dataset T with a sample size of M is obtained. M *, where the peak level of the new data can be adjusted by probability πj. When πj = 1 / 2, i = 0, 1, the sample sizes of the two types of data are roughly the same.

[0060] In order to balance the training data set data while retaining the original data without the data information of the lung infection group, the training data set in the original data was processed by the ROSE algorithm to generate a new training data set of 881÷0.5=1762 cases, of which the lung infection group data amounted to 845 cases and the non-lung infection group data amounted to 917 cases. This part of the processed data is defined as artificial data. In the study of prediction models, the purpose of using the ROSE (Random Over-SamplingExamples) algorithm is to solve the problem of class imbalance. Specifically, the ROSE algorithm is used to deal with the situation of class imbalance in training data, in which the number of samples in one category is much less than the number of samples in another category. The ROSE algorithm balances the sample distribution between different categories by synthesizing new minority class samples, thereby improving the model's recognition ability for minority class samples and enhancing the model's generalization ability and prediction performance. The application of the ROSE algorithm can help improve the performance of the model on unbalanced data sets and improve the prediction accuracy of the model.

[0061] Data analysis showed that age, white blood cell count (WBC), and albumin-globulin ratio (AGR) could be used as predictors of postoperative pulmonary infection in elderly patients with hip fractures, i.e., 1341 patients were randomly divided into a training data set of 939 cases and a test data set of 402 cases (random seed was 1122) in a ratio of 7:3. Literature confirmed that gender and hemoglobin (HGB) were also independent risk factors for postoperative pulmonary infection, and they were objective indicators, and relevant data could be obtained before the patient's surgery; therefore, based on the statistical analysis results and clinical practical application considerations, the five predictive variables finally included in the model were determined: age, gender, WBC, AGR, and HGB.

[0062] The modeling sample size should meet the requirement of 10 patients with postoperative lung infection for each risk factor. This application randomly divides 1341 patients into a training data set of 939 cases and a test data set of 402 cases in a ratio of 7:3 (the random seed is 1122); the number of cases in the lung infection group in the training data set is 58, and the number of cases in the lung infection group in the test data set is 21. The incidence of postoperative lung infection in the training data set is 6.2%, and there are 5 variables finally included in the model. Therefore, the sample size of the modeling group should be at least 5×10÷0.062≈807<939, so the training data set of this application meets the modeling sample size requirements.

[0063] The comparison of clinical characteristics between the training data set and the test data set showed that there were 939 cases in the original data, including 58 patients with postoperative pulmonary infection (6.2%) in the training data set, and 402 cases in the test data set, including 21 patients with postoperative pulmonary infection (5.2%). There was no statistically significant difference in age, gender, preoperative comorbidity-related indicators, laboratory-related indicators, ICU admission rate, and in-hospital death between the two groups (P>0.05), indicating that the two groups were comparable.

[0064] (1) Construction and internal validation of a prediction model for pulmonary infection after hip fracture surgery in the elderly.

[0065] Based on the training set data of 1762 cases in the artificial data, age, gender, WBC, AGR and HGB were included to establish a prediction model based on binary logistic regression. The model formula was Z = 0.034 × age - 0.416 × gender + 0.094 × WBC - 1.240 × AGR - 0.014 × HGB - 0.103 (male = 0, female = 1), and the risk probability P = 1 / (1 + eZ). The area under the receiver operating characteristic curve (ROC) of the modeling group was 0.699, the sensitivity was 63.7%, the specificity was 66.5%, and the accuracy was 65.1%. The internal validation was performed by internal validation of the artificial data after the ROSE algorithm in the early stage through the grouped 402 cases of data. The results showed that the AUC value of the model was 0.781, the sensitivity was 92.4%, the specificity was 57.1%, and the accuracy was 90.5%.

[0066] (2) External validation of the prediction model for pulmonary infection after hip fracture surgery in the elderly.

[0067] In order to verify the extrapolation and universality of the previous model and make the model truly applicable to clinical practice, we also collected clinical data from elderly patients with hip fractures from three hospitals to externally verify and update the previous model. Among them, Shenzhen Second People's Hospital collected 411 cases from January 2021 to June 2023, Shenzhen Pingle Orthopedic Hospital collected 371 cases from January 2021 to December 2022, and Peking University Shenzhen Hospital collected 194 cases from January 2021 to December 2022. A total of 976 elderly patients with hip fractures were collected for external verification and update of the model.

[0068] The predictive factors (age, gender, white blood cells, white blood cell ratio and hemoglobin) of the prediction model established earlier were put into the Logistic regression model with whether the patient had lung infection as the result. The results showed that the AUC value of the model was 0.726 (95% CI: 0.664-0.784), the sensitivity was 81.3%, the specificity was 55.0%, and the accuracy was 56.8%. The calibration curve results showed that the slope of the calibration curve was 0.305, and its calibration was poor, so the model needed to be updated.

[0069] The methods for optimizing clinical prediction models mainly include regression coefficient updating, Meta-model method and dynamic model construction. Since Meta-model method and dynamic model construction have many limitations in practical applications, regression coefficient updating is currently the most commonly used method to optimize the model. However, although the sample size included in this study is slightly higher than the modeling cohort, the sample size of the cohort in each center is relatively small. In order to ensure that the model is less affected by the sample size, this study updates and optimizes the intercept and regression coefficient based on the original model.

[0070] ① Only correct the intercept: Model 1.

[0071] The corrected model is Model 1 = 0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB+α1=-2.538+0.034*age-0.416*sex+0.094*WBC1.240*AGR-0.014*HGB; among which, α1=-2.538.

[0072] ② Correct the intercept and regression coefficient simultaneously: Model 2.

[0073] The corrected model is Model 2 = α2 + β*(0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB) = -2.766+0.042*age-0.515*sex+0.116*WBC-1.534*AGR-0.017*HGB; among which, α2 = -2.766, β = 1.23726.

[0074] After calibration of the two methods, the AUC value of the two models was 0.726, the sensitivity was 81.3%, the specificity was 55.0%, and the accuracy was 56.8%. The calibration curves of Model 1 and Model 2 showed slopes of 0.938 and 0.858, respectively, both close to 1.000, and good calibration. The optimal cutoff values ​​were 0.066 and 0.052, respectively.

[0075] (3) Model training.

[0076] For example, when a male patient is 85 years old, with a preoperative white blood cell count of 11.9*10^9 / L, an albumin-globulin ratio of 2.0, and a hemoglobin level of 115g / L, when Model 1 is used to calculate the probability of postoperative lung infection in this patient, it is 0.068, which is higher than the optimal risk threshold of 0.066, so this patient is considered to be at high risk of lung infection; when Model 2 is used to calculate the probability of postoperative lung infection in this patient, it is 0.055, which is higher than the optimal risk threshold of 0.052 of its model, so it is also considered to be at high risk of postoperative lung infection. Although Model 1 and Model 2 predict different probabilities of postoperative lung infection for the same patient, because the discrimination of the two models is consistent, the models are consistent in predicting whether the patient is a high-risk group.

[0077] Clinical prediction models, also known as "prognostic models" or "risk scores," are used to estimate the risk of developing a specific disease and stratify patients based on this risk. It does this by assigning a corresponding weight to each predictive factor in order to more accurately predict the risk a patient may face.

[0078] A risk prediction calculator that can predict the probability of disease occurrence is constructed through a clinical prediction model, which can provide a convenient risk prediction tool for clinicians and achieve the following goals:

[0079] (1) Provide decision support; ② Optimize resource utilization; ③ Prevention and early intervention; ④ Improve patient safety; ⑤ Personalized medicine; ⑥ Manage chronic diseases.

[0080] Previous studies have established a risk prediction model for postoperative pulmonary infection in patients with hip fractures based on objective laboratory test indicators and artificial data. Laboratory test indicators are the material basis for the body to maintain balance in all aspects. Abnormal results often appear before clinical symptoms appear and are easy to obtain, which has the value of early diagnosis and prediction of diseases.

[0081] In order to further evaluate the generalizability of the model, the prediction model was externally validated. This study used different time periods and added two more hospitals as research sites on the basis of the original research sites. Therefore, this study mainly adopted the external validation method of time period + space, which solved the problem that internal validation could not fully reflect the generalization ability of the model. Through external validation at different times and places, according to the requirements of the modeling cohort, the new discrimination and calibration results of the model after the predictive factors of the model were fitted on the new data set were achieved, realizing the key step of transforming the clinical prediction model from theory to clinical practice.

[0082] like Figure 2-Figure 3 As shown. Figure 2-Figure 3 Two prediction models are presented, both of which predict the risk of postoperative pulmonary infection in elderly patients with hip fractures. Model 1 is obtained by updating the intercept of the original model, and Model 2 is obtained by updating the intercept and regression coefficient of the original model at the same time.

[0083] The present application can further improve the prediction accuracy of the lung infection risk probability by calibrating the lung infection risk probability prediction model.

[0084] In an exemplary embodiment, a computer device is provided, which may be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Wherein, the processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. Wherein, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for predicting the probability of a risk of pulmonary infection is implemented.

[0085] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0086] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0087] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0088] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0089] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0090] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0091] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for predicting the probability of lung infection risk, characterized in that: The method for predicting the probability of lung infection risk comprises: Obtaining a training data set; the training data set includes multiple groups of data, each group of data includes age, gender, white blood cell count, albumin-globulin ratio, hemoglobin, and actual risk probability value of lung infection; Based on the training data set, determining whether the predicted probability value of the lung infection risk of the lung infection risk probability prediction model and the actual risk probability value are within a set error range; the lung infection risk probability prediction model is established with the age, the gender, the white blood cell count, the albumin-globulin ratio and the hemoglobin as inputs and the actual lung infection risk probability value as output; If yes, predicting the lung infection risk probability value of the postoperative patient according to the lung infection risk probability prediction model; If not, calibrating the lung infection risk probability prediction model; The lung infection risk probability value of postoperative patients was predicted based on the corrected lung infection risk probability prediction model.

2. The method for predicting the probability of lung infection risk according to claim 1, characterized in that: The lung infection risk probability prediction model is corrected by using an intercept correction method, including: Based on the training data set, a correction factor is calculated according to the predicted probability value of the lung infection risk of the lung infection risk probability prediction model and the actual probability value of the lung infection risk; Determining a corrected intercept according to the correction factor and the intercept of the lung infection risk probability prediction model; The lung infection risk probability prediction model is corrected according to the corrected intercept, and a corrected lung infection risk probability prediction model is established.

3. The method for predicting the probability of lung infection risk according to claim 2, characterized in that: The correction factor=Ln((actual risk probability value of lung infection / (1-actual risk probability value of lung infection)) / (predicted risk probability value of lung infection / (1-predicted risk probability value of lung infection))).

4. The method for predicting the probability of lung infection risk according to claim 3, characterized in that: The corrected intercept α1=α0+correction factor; wherein α0 is the intercept of the lung infection risk probability prediction model.

5. The method for predicting the probability of lung infection risk according to claim 4, characterized in that: The corrected prediction model for the probability of lung infection risk was: Model 1 = 0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB+α1; Among them, Model 1 is the adjusted lung infection risk probability prediction model, which is a prediction model corrected by intercept correction; WBC is white blood cell count; AGR is albumin-globulin ratio; HGB is hemoglobin.

6. The method for predicting the probability of lung infection risk according to claim 1, characterized in that: The lung infection risk probability prediction model is corrected by using the intercept and regression coefficient simultaneous correction method, including: Updating the intercept of the lung infection risk probability prediction model to generate an updated intercept; Based on the intercept of the lung infection risk probability prediction model, the regression coefficients of all predictors are updated as a whole; the predictors include age, gender, white blood cell count, albumin-globulin ratio and hemoglobin; The lung infection risk probability prediction model is corrected according to the updated intercept and the overall updated regression coefficient to establish a corrected lung infection risk probability prediction model.

7. The method for predicting the probability of lung infection risk according to claim 6, characterized in that: The corrected prediction model for the probability of lung infection risk was: Model 2 = α2 + β*(0.034*age-0.416*sex+0.094*WBC-1.240*AGR-0.014*HGB); Among them, Model 1 is the corrected lung infection risk probability prediction model, which is a prediction model calibrated by simultaneously correcting the intercept and regression coefficient; α2 is the updated intercept; β is the overall updated regression coefficient; WBC is the white blood cell count; AGR is the albumin-globulin ratio; HGB is hemoglobin.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting the probability of lung infection risk according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the probability of lung infection risk according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting the probability of lung infection risk according to any one of claims 1 to 6 is implemented.

Images