Death risk prediction model for severe burn combined with hyperbilirubinemia, construction method and storage medium

Abstract

The invention relates to a death risk prediction model for severe burns and hyperbilirubinemia, a construction method and a storage medium. The construction method comprises the following steps: collecting clinical feature data of patients with severe burns and hyperbilirubinemia; in combination with LASSO and Logistic regression, screening out clinical feature data associated with severe burn and hyperbilirubinemia from the clinical feature data; and constructing a death risk prediction model based on the screened clinical feature data. The invention also provides a death risk prediction model for severe burn combined with hyperbilirubinemia, which is constructed by the construction method. The invention further provides a computer readable storage medium. According to the method, risk factors causing death of patients suffering from severe burns and hyperbilirubinemia abnormality are identified based on clinical basic data and disease related indexes of the patients, so that the death probability of the patients is conveniently, quickly and accurately predicted in advance, and doctors are helped to take corresponding prevention and treatment measures for the patients.

Description

Technical Field

[0001] The present invention relates to the field of medical diagnosis technology, and in particular to a death risk prediction model for severe burns combined with hyperbilirubinemia, a construction method and a storage medium. Background Art

[0002] Burns are a very common type of trauma in daily life, which poses a great threat to human life and health. Burns may not only cause long-term physical and mental problems, but also a large number of patients lose their precious lives due to burns every year around the world. Burns are classified mainly based on the depth of the burn and the area of ​​the burn. The depth of burns ranges from first to fourth degree, while the area of ​​the burn is divided into mild and severe. Mild burns are characterized by a burn area of ​​less than 10% of the total body surface area (TBSA), mainly manifested as surface burns. In general, if the burn area exceeds 30% TBSA in children, 20% TBSA in adults, and 10% TBSA in elderly patients, this is usually considered a severe burn.

[0003] Severe burns may induce strong and persistent systemic reactions. After a burn occurs, the systemic inflammatory response aimed at promoting healing may be over-amplified and become uncontrolled. This aggravated inflammatory response will gradually lead to a systemic hypermetabolic state and delayed wound healing, which in turn leads to poor prognosis, organ dysfunction, infection, and even death. Currently, many studies have focused on the therapeutic effects of infection, blood transfusion, allogeneic skin transplantation, initial surgical management, and new combined therapies on burn-related multiple organ injury (MBI), but often ignore the impact of other factors on the occurrence and progression of the disease, which limits the effective guidance of early intervention of the disease. Studies have shown that if burn patients have abnormal bilirubin, their mortality rate will increase significantly, but it is not clear which specific indicators will affect the mortality of these patients. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a death risk prediction model, construction method and storage medium for severe burns combined with hyperbilirubinemia, so as to identify the risk factors leading to death of patients with severe burns and abnormal hyperbilirubinemia based on the patient's clinical basic data and disease-related indicators, and use these factors to establish a fully verified death risk prediction model to predict the patient's probability of death in advance, conveniently, quickly and accurately, thereby helping doctors take corresponding preventive and treatment measures for patients.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia comprises the following steps:

[0007] S1. Data collection and preprocessing: Clinical characteristic data of patients with severe burns and hyperbilirubinemia were collected and preprocessed;

[0008] S2. Screening of clinical characteristic data: First, LASSO regression was used to screen the pre-processed clinical characteristic data, and clinical characteristic data associated with severe burns combined with hyperbilirubinemia were preliminarily screened out. The preliminarily screened clinical characteristic data were divided into a training set and a validation set. The clinical characteristic data of the training set were used to screen out clinical characteristic data with p less than 0.05 and OR value greater than 1 through Logistic regression as clinical characteristic data associated with severe burns combined with hyperbilirubinemia and death;

[0009] S3. Model construction: Based on the screened indicators, a death risk prediction model is constructed.

[0010] According to the above technical means, the quality and relevance of the data were ensured by processing the collected data. LASSO regression was used for preliminary screening, which effectively reduced the feature dimensions while retaining the features related to the target variable, improving the predictive performance of the model. Dividing the clinical feature data into training sets and validation sets helps to evaluate the generalization ability of the model and reduce the risk of overfitting. Logistic regression further screened out statistically significant clinical features to ensure the clinical relevance and interpretability of the model. The constructed death risk prediction model can provide strong support for clinical decision-making and help improve the treatment effect and survival rate of patients with severe burns and hyperbilirubinemia.

[0011] Among them, clinical data features with non-zero regression coefficients screened out using LASSO regression were considered to be clinical characteristic data associated with severe burns combined with hyperbilirubinemia.

[0012] Preferably, the clinical characteristic data include: age, gender, BMI, superficial second-degree burn area, deep second-degree burn area, full-thickness burn area, total burn area, burn index (BI), acute burn severity index, Baux score, whether it is flame burn, chemical burn, electrical burn, high-temperature gas burn, explosion injury, arc burn, hot liquid burn, scald, solid burn, inhalation injury, history of diabetes, history of tumor, history of heart disease, history of lung disease, and whether there is self-injury. History of personal immune diseases, history of infectious diseases, history of elevated blood lipids, albumin on admission, albumin before admission, albumin-globulin ratio on admission, alanine aminotransferase on admission, aspartate aminotransferase on admission, glutamyl transferase on admission, alkaline phosphatase on admission, lactate dehydrogenase on admission, direct bilirubin on admission, indirect bilirubin on admission, total bilirubin on admission, fibrinogen on admission, blood creatinine on admission, maximum value of alanine aminotransferase, maximum value of aspartate aminotransferase, maximum value of glutamyl transferase, maximum value of alkaline phosphatase, maximum value of direct bilirubin, maximum value of indirect bilirubin and maximum value of total bilirubin.

[0013] Preferably, the clinical characteristic data related to death of severe burns combined with hyperbilirubinemia include: age, burn index (BI), admission serum creatinine and maximum direct bilirubin.

[0014] Preferably, the mortality risk prediction model is a nomogram.

[0015] Preferably, in the nomogram:

[0016] The first column is the score scale, ranging from 0 to 100;

[0017] The second column is the patient's age, including age scores, including 0 to 10 years old, corresponding to 0; 10 to 20 years old, corresponding to 0 to 12; 20 to 30 years old, corresponding to 12 to 25; 30 to 40 years old, corresponding to 25 to 38; 40 to 50 years old, corresponding to 38 to 50; 50 to 60 years old, corresponding to 50 to 62; 60 to 70 years old, corresponding to 62 to 75; 70 to 80 years old, corresponding to 75 to 88; 80 to 90 years old, corresponding to 88 to 100;

[0018] The third column is the patient's admission serum creatinine, and the admission serum creatinine score includes 0 mg / dL corresponding to 0; 0-50 mg / dL corresponding to 0-9; 50-150 mg / dL corresponding to 9-17; 150-200 mg / dL corresponding to 17-26; 200-250 mg / dL corresponding to 26-44; 250-300 mg / dL corresponding to 44-52; 300-350 mg / dL corresponding to 52-61;

[0019] 350-400 mg / dL corresponds to a score of 61-70; 400-450 mg / dL corresponds to a score of 70-79; 450-500 mg / dL corresponds to a score of 79-87;

[0020] The fourth column is the maximum direct bilirubin value, and the maximum direct bilirubin value score includes 0 mg / dL corresponding to a score of 0;

[0021] 0-20 mg / dL corresponds to a score of 0-8; 20-40 mg / dL corresponds to a score of 8-16; 40-60 mg / dL corresponds to a score of 16-23;

[0022] 60-80 mg / dL corresponds to a score of 23-31; 80-100 mg / dL corresponds to a score of 31-39; 100-120 mg / dL corresponds to a score of 39-47; 120-140 mg / dL corresponds to a score of 47-54; 140-160 mg / dL corresponds to a score of 54-62; 160-180 mg / dL corresponds to a score of 62-70; 180-200 mg / dL corresponds to a score of 70-78;

[0023] The fifth column is the burn index (BI), and the BI score includes 0 mg / dL corresponding to 0; 0-10 mg / dL corresponding to 0-9; 10-20 mg / dL corresponding to 9-17; 20-30 mg / dL corresponding to 17-26; 30-40 mg / dL corresponding to 26-34; 40-50 mg / dL corresponding to 34-43; 50-60 mg / dL corresponding to 43-52; 60-70 mg / dL corresponding to 52-60; 70-80 mg / dL corresponding to 60-69; 80-90 mg / dL corresponding to 69-77; 90-100 mg / dL corresponding to 77-86; the sixth column is the total score; the seventh column is the probability of death risk.

[0024] Among them, the second, third, fourth and fifth columns are related factors, and different situations of different factors correspond to different scores of the scale. The sixth and seventh columns are corresponding relationships, and different total scores correspond to different predicted values ​​of the probability of death risk. The higher the patient's total score, the higher the risk of death in severe burns combined with hyperbilirubinemia.

[0025] Preferably, the AUC of the nomogram is 0.882.

[0026] Preferably, the death risk probability calculation formula of the death risk prediction model is:

[0027]

[0028] In formula (I), X1 represents age; X2 represents the blood creatinine value at admission; X3 represents the maximum value of direct bilirubin; X4 represents the burn index (BI); and P represents the mortality risk probability of severe burns combined with hyperbilirubinemia.

[0029] Preferably, the method for constructing the mortality risk prediction model for severe burns combined with hyperbilirubinemia comprises:

[0030] The clinical characteristic data initially screened were divided into a 70% training set and a 30% validation set; the model was trained using the training data set to construct a mortality risk prediction model;

[0031] The death risk prediction model was validated using the validation set.

[0032] The present invention also provides a mortality risk prediction model for severe burns combined with hyperbilirubinemia constructed by the construction method of the present invention.

[0033] Preferably, the mortality risk prediction model for severe burns combined with hyperbilirubinemia includes:

[0034] Data collection and preprocessing module, used to collect clinical characteristic data of patients with severe burns and hyperbilirubinemia;

[0035] The clinical characteristic data screening module is used to first use LASSO regression to screen the preprocessed clinical characteristic data, preliminarily screen out the clinical characteristic data associated with severe burns combined with hyperbilirubinemia, and then use Logistic regression to screen out the clinical characteristic data with p less than 0.05 and OR value greater than 1 as the clinical characteristic data related to severe burns combined with hyperbilirubinemia and death;

[0036] The prediction model building module constructs a death risk prediction model based on the screened clinical characteristic data related to severe burns combined with hyperbilirubinemia and death.

[0037] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the construction method described in the present invention are implemented.

[0038] Beneficial effects of the present invention:

[0039] The method for constructing a death risk prediction model for severe burns combined with hyperbilirubinemia of the present invention ensures that the basic data for model construction is rich and comprehensive by collecting clinical characteristic data of 47 variables including age, admission blood creatinine, BI and maximum direct bilirubin, which helps to improve the accuracy of prediction; the combination of LASSO regression and Logistic regression two screening methods can more comprehensively identify clinical characteristic data related to patient mortality and reduce the risk of missing important variables; LASSO regression has the characteristics of automatic variable selection and regularization, which helps to process high-dimensional data and prevent overfitting; Logistic regression can provide the statistical significance of variables, ensuring that the selected variables have a statistically significant impact on the prediction results; the death risk prediction model constructed based on the screened indicators can provide clinicians with an accurate and practical tool, realize result visualization, help evaluate the patient's death risk, and thus formulate a more personalized treatment plan, which has promotion and application value in the field of medical diagnostic technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A nomogram model for predicting mortality risk in patients with severe burns and hyperbilirubinemia;

[0041] Figure 2 This is the ROC curve of the validation set;

[0042] Figure 3 DAC curve diagram for the validation set;

[0043] Figure 4 This is the calibration curve of the validation set.

[0044] Figure 5 It is the ROC curve of the training set;

[0045] Figure 6 DAC curve diagram for the training set;

[0046] Figure 7 is the calibration curve diagram of the training set;

[0047] Figure 8 The ROC curve diagram for the training set and the validation set. DETAILED DESCRIPTION

[0048] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0049] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0050] The present invention intends to disclose a death risk prediction model for severe burns combined with hyperbilirubinemia, a construction method and a storage medium, so as to identify the risk factors leading to death of patients with severe burns and abnormal hyperbilirubinemia based on the patient's clinical basic data and disease-related indicators, and use these factors to establish a fully verified death risk prediction model to conveniently, quickly and accurately predict the patient's probability of death in advance, thereby helping doctors take corresponding preventive and treatment measures for the patient.

[0051] Among them, the method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia includes the following steps:

[0052] S1. Data collection and preprocessing: Clinical characteristics data of patients with severe burns and hyperbilirubinemia were collected;

[0053] S2. Index screening: First, LASSO regression was used to screen the pre-processed clinical characteristic data, and clinical characteristic data associated with severe burns combined with hyperbilirubinemia were preliminarily screened out. The preliminarily screened clinical characteristic data were divided into a training set and a validation set. The clinical characteristic data of the training set were used to screen out clinical characteristic data with p less than 0.05 and OR value greater than 1 through Logistic regression as clinical characteristic data associated with severe burns combined with hyperbilirubinemia and death;

[0054] S3. Model construction: Based on the screened clinical characteristic data related to severe burns combined with hyperbilirubinemia and death, a death risk prediction model was constructed.

[0055] Among them, the model in S3 was constructed using R language, combining R language's ggplot2, lattice, Hmisc, openxlsx, epiDisplay, foreign, survival, MASS and nnet logical packages to construct a mortality risk prediction model.

[0056] In some embodiments, the clinical characteristic data include: age, gender, BMI, superficial second-degree burn area, deep second-degree burn area, full-thickness burn area, total burn area, burn index (BI), acute burn severity index, Baux score, whether it is a flame burn, whether it is a chemical burn, whether it is an electrical burn, whether it is a high-temperature gas burn, whether it is an explosion injury, whether it is an arc burn, whether it is a hot liquid burn, whether it is a scald, whether it is a solid burn, whether it is an inhalation injury, whether there is a history of diabetes, whether there is a history of tumors, whether there is a history of heart disease, whether there is a history of lung disease, whether there is History of autoimmune disease, history of infectious disease, history of elevated blood lipids, albumin on admission, albumin before admission, albumin-globulin ratio on admission, alanine aminotransferase on admission, aspartate aminotransferase on admission, glutamyl transferase on admission, alkaline phosphatase on admission, lactate dehydrogenase on admission, direct bilirubin on admission, indirect bilirubin on admission, total bilirubin on admission, fibrinogen on admission, blood creatinine on admission, maximum value of alanine aminotransferase, maximum value of aspartate aminotransferase, maximum value of glutamyl transferase, maximum value of alkaline phosphatase, maximum value of direct bilirubin, maximum value of indirect bilirubin and maximum value of total bilirubin.

[0057] In some embodiments, clinical characteristic data related to death in severe burns combined with hyperbilirubinemia screened out by combining LASSO regression and Logistic regression include: age, burn index (BI), admission serum creatinine, and maximum direct bilirubin.

[0058] In some embodiments, the mortality risk prediction model is a nomogram.

[0059] Exemplarily, in S1, data collection and preprocessing include:

[0060] Data of 321 patients with severe burns and hyperbilirubinemia who were consecutively admitted to the Burn Department of Southwest Hospital from January 2017 to December 2022 were collected and pre-processed. The pre-processing specifically excluded patients in the following situations: (1) minors, defined as individuals under the age of 18; (2) liver damage caused by reasons other than burns; (3) hospitalization time less than 24 hours; (4) patients admitted more than 7 days after burns; (5) patients with major missing data; (6) patients whose biochemical parameters were not measured within 72 hours of admission; (7) patients with a burn area less than 20%. Among them, patients with major missing data generally include patients who only have a few basic information about admission, such as name, age, admission time, etc., but no biochemical data or various examination indicators.

[0061] In some embodiments, Figure 1 As shown in the nomogram:

[0062] The first column is the score scale, ranging from 0 to 100;

[0063] The second column is the patient's age, including age scores, including 0 to 10 years old, corresponding to 0; 10 to 20 years old, corresponding to 0 to 12; 20 to 30 years old, corresponding to 12 to 25; 30 to 40 years old, corresponding to 25 to 38; 40 to 50 years old, corresponding to 38 to 50; 50 to 60 years old, corresponding to 50 to 62; 60 to 70 years old, corresponding to 62 to 75; 70 to 80 years old, corresponding to 75 to 88; 80 to 90 years old, corresponding to 88 to 100;

[0064] The third column is the patient's admission serum creatinine, and the admission serum creatinine score includes 0 mg / dL corresponding to 0; 0-50 mg / dL corresponding to 0-9; 50-150 mg / dL corresponding to 9-17; 150-200 mg / dL corresponding to 17-26; 200-250 mg / dL corresponding to 26-44; 250-300 mg / dL corresponding to 44-52; 300-350 mg / dL corresponding to 52-61;

[0065] 350-400 mg / dL corresponds to a score of 61-70; 400-450 mg / dL corresponds to a score of 70-79; 450-500 mg / dL corresponds to a score of 79-87;

[0066] The fourth column is the maximum direct bilirubin value, and the maximum direct bilirubin value score includes 0 mg / dL corresponding to a score of 0;

[0067] 0-20 mg / dL corresponds to a score of 0-8; 20-40 mg / dL corresponds to a score of 8-16; 40-60 mg / dL corresponds to a score of 16-23;

[0068] 60-80 mg / dL corresponds to a score of 23-31; 80-100 mg / dL corresponds to a score of 31-39; 100-120 mg / dL corresponds to a score of 39-47; 120-140 mg / dL corresponds to a score of 47-54; 140-160 mg / dL corresponds to a score of 54-62; 160-180 mg / dL corresponds to a score of 62-70; 180-200 mg / dL corresponds to a score of 70-78;

[0069] The fifth column is the burn index (BI), and the BI score includes 0 mg / dL corresponding to 0; 0-10 mg / dL corresponding to 0-9; 10-20 mg / dL corresponding to 9-17; 20-30 mg / dL corresponding to 17-26; 30-40 mg / dL corresponding to 26-34; 40-50 mg / dL corresponding to 34-43; 50-60 mg / dL corresponding to 43-52; 60-70 mg / dL corresponding to 52-60; 70-80 mg / dL corresponding to 60-69; 80-90 mg / dL corresponding to 69-77; 90-100 mg / dL corresponding to 77-86; the sixth column is the total score; the seventh column is the probability of death risk.

[0070] The second, third, fourth and fifth columns are related factors, and different factors correspond to different scores on the scale. The sixth and seventh columns are corresponding relationships, and different total scores correspond to different predicted values ​​of the probability of death risk. Figure 1 As shown in the data, the higher the total score of the patient, the higher the risk of death in severe burns with hyperbilirubinemia.

[0071] In some embodiments, the death risk probability calculation formula of the death risk prediction model is:

[0072]

[0073] In formula (I), X1 represents age; X2 represents the blood creatinine value at admission; X3 represents the maximum value of direct bilirubin; X4 represents the burn index (BI); and P represents the mortality risk probability of severe burns combined with hyperbilirubinemia.

[0074] In some embodiments, a method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia comprises:

[0075] The clinical characteristic data preliminarily screened were divided into a 70% training set and a 30% validation set; the model was trained using the training data set to construct a mortality risk prediction model;

[0076] The mortality risk prediction model was internally validated using the validation set.

[0077] Specifically: The specific operation method of internal validation is: the clinical characteristic data of patients with severe burns and hyperbilirubinemia were collected and screened through LASSO regression, and the clinical characteristic data were screened out, including age, superficial second-degree burn area, admission lactate dehydrogenase, admission blood creatinine, maximum direct bilirubin, acute burn severity index, burn index, whether it is electric current burn, whether it is hydrothermal burn, and history of heart disease; then the patient data of these clinical characteristic data were divided into 70% training set and 30% validation set, and the training set was used to build a model, and Logistic screening was performed to screen out 4 variables, and a model was built for them, visualized as a nomogram, and then the validation set was used for internal validation.

[0078] Among them, the ROC curve of the validation set is as follows Figure 2 As shown. Figure 2 From the analysis, we can see that the area under the ROC curve (AUC) reaches 0.882, close to 0.9, and the prediction effect is good.

[0079] The DAC curve of the validation set is as follows: Figure 3 shown. Figure 3 In the figure, the solid line is the baseline, and the DAC curve is above the baseline throughout, indicating that the model works well.

[0080] The calibration curve of the validation set is as follows Figure 4 shown. Figure 4 The apparent curve in is close to the ideal curve, and the prediction effect of the model is good.

[0081] The above validation results show that the internal validation results show good prediction effect.

[0082] Among them, the ROC curve of the training set is as follows Figure 5 The DAC curve of the training set is shown in Figure 6 The calibration curve of the training set is shown in Figure 7 The total ROC curve of the training set and the validation set is shown in Figure 8 shown.

[0083] In some embodiments, a mortality risk prediction model for severe burns combined with hyperbilirubinemia constructed by the construction method in any of the above embodiments is also provided.

[0084] In some embodiments, the mortality risk prediction model for severe burns combined with hyperbilirubinemia comprises:

[0085] Data collection and preprocessing module, used to collect clinical characteristic data of patients with severe burns and hyperbilirubinemia;

[0086] The clinical characteristic data screening module is used to first use LASSO regression to screen the pre-processed clinical characteristic data, preliminarily screen out the clinical characteristic data associated with severe burns combined with hyperbilirubinemia, divide the preliminarily screened clinical characteristic data into a training set and a validation set, and use the clinical characteristic data of the training set to screen out the clinical characteristic data with p less than 0.05 and or value greater than 1 through Logistic regression as the clinical characteristic data related to severe burns combined with hyperbilirubinemia and death;

[0087] The prediction model building module constructs a death risk prediction model based on the screened clinical characteristic data related to severe burns combined with hyperbilirubinemia and death.

[0088] In some embodiments, the mortality risk prediction model for severe burns combined with hyperbilirubinemia further includes:

[0089] A storage module stores various information during the operation of the death risk prediction model for severe burns combined with hyperbilirubinemia;

[0090] Communication module, used for communication connection between modules in the model and between the prediction model and external devices;

[0091] The control module controls the normal operation of the input data collection and preprocessing module, the clinical characteristic data screening module, the prediction model building module, the storage module and the communication module.

[0092] In some embodiments, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the construction method in any of the above embodiments are implemented.

[0093] In summary, the method for constructing a death risk prediction model for severe burns combined with hyperbilirubinemia of the present invention ensures that the basic data for model construction is rich and comprehensive by collecting clinical characteristic data of 47 variables including age, admission blood creatinine, BI and maximum direct bilirubin, which helps to improve the accuracy of prediction; the combination of LASSO regression and Logistic regression two screening methods can more comprehensively identify clinical characteristic data related to patient mortality and reduce the risk of missing important variables; LASSO regression has the characteristics of automatic variable selection and regularization, which helps to process high-dimensional data and prevent overfitting; Logistic regression can provide the statistical significance of variables, ensuring that the selected variables have a statistically significant impact on the prediction results; the death risk prediction model constructed based on the screened indicators can provide clinicians with an accurate and practical tool, realize result visualization, help assess the patient's death risk, and thus formulate a more personalized treatment plan, which has promotion and application value in the field of medical diagnostic technology.

[0094] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia, characterized in that: The following steps are involved: S1. Data collection and preprocessing: Clinical characteristic data of patients with severe burns and hyperbilirubinemia were collected and preprocessed; S2. Screening of clinical characteristic data: First, LASSO regression was used to screen the pre-processed clinical characteristic data, and clinical characteristic data associated with severe burns combined with hyperbilirubinemia were preliminarily screened out. The preliminarily screened clinical characteristic data were divided into a training set and a validation set. The clinical characteristic data of the training set were used to screen out clinical characteristic data with p less than 0.05 and OR value greater than 1 through Logistic regression as clinical characteristic data associated with severe burns combined with hyperbilirubinemia and death; S3. Model construction: Based on the screened clinical characteristic data related to severe burns combined with hyperbilirubinemia and death, a death risk prediction model was constructed.

2. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 1, characterized in that: The clinical characteristic data include: age, gender, BMI, superficial second-degree burn area, deep second-degree burn area, full-thickness burn area, total burn area, burn index (BI), acute burn severity index, Baux score, whether it is flame burn, chemical burn, electrical burn, high-temperature gas burn, explosion injury, arc burn, hot liquid burn, scald, solid burn, inhalation injury, history of diabetes, history of tumor, history of heart disease, history of lung disease, autoimmune disease Medical history, whether there is a history of infectious diseases, whether there is a history of elevated blood lipids, admission albumin, pre-admission albumin, admission albumin-globulin ratio, admission alanine aminotransferase, admission aspartate aminotransferase, admission glutamyl transferase, admission alkaline phosphatase, admission lactate dehydrogenase, admission direct bilirubin, admission indirect bilirubin, admission total bilirubin, admission fibrinogen, admission blood creatinine, maximum alanine aminotransferase, maximum aspartate aminotransferase, maximum glutamyl transferase, maximum alkaline phosphatase, maximum direct bilirubin, maximum indirect bilirubin and maximum total bilirubin; And / or, the clinical characteristic data related to death of severe burns combined with hyperbilirubinemia, including: age, burn index (BI), admission serum creatinine and maximum direct bilirubin.

3. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 1, characterized in that: The death risk prediction model is a nomogram.

4. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 3, characterized in that: In the nomogram: The first column is the score scale, with a score range of 0 to 100; the second column is the patient's age; the third column is the patient's admission serum creatinine; the fourth column is the maximum direct bilirubin; the fifth column is the burn index (BI); the sixth column is the total score; and the seventh column is the probability of death risk.

5. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 4, characterized in that: The nomogram had an AUC = 0.

882.

6. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 1, characterized in that: The death risk probability calculation formula of the death risk prediction model is: In formula (I), X1 represents age; X2 represents the blood creatinine value at admission; X3 represents the maximum value of direct bilirubin; X4 represents the burn index (BI); and P represents the mortality risk probability of severe burns combined with hyperbilirubinemia.

7. The method for constructing a mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 1, characterized in that: The method for constructing the death risk prediction model for severe burns combined with hyperbilirubinemia comprises: The clinical characteristic data preliminarily screened were divided into a 70% training set and a 30% validation set; the model was trained using the training data set to construct a mortality risk prediction model; The death risk prediction model was validated using the validation set.

8. A mortality risk prediction model for severe burns combined with hyperbilirubinemia constructed according to the construction method according to any one of claims 1 to 7.

9. The mortality risk prediction model for severe burns combined with hyperbilirubinemia according to claim 8, characterized in that: include: Data collection and preprocessing module, used to collect clinical characteristic data of patients with severe burns and hyperbilirubinemia; The clinical characteristic data screening module is used to first use LASSO regression to screen the pre-processed clinical characteristic data, preliminarily screen out the clinical characteristic data associated with severe burns combined with hyperbilirubinemia, divide the preliminarily screened clinical characteristic data into a training set and a validation set, and use the clinical characteristic data of the training set to screen out the clinical characteristic data with p less than 0.05 and or value greater than 1 through Logistic regression as the clinical characteristic data related to severe burns combined with hyperbilirubinemia and death; The prediction model building module constructs a death risk prediction model based on the screened clinical characteristic data related to severe burns combined with hyperbilirubinemia and death.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the construction method according to any one of claims 1 to 7 are implemented.

Images