Construction method of acute mushroom poisoning death prediction model

Abstract

The invention relates to a construction method of an acute mushroom poisoning death prediction model, which comprises the following steps: collecting detection data of a plurality of acute mushroom poisoning persons, and constructing a training set and a verification set; performing single-factor analysis based on the training set to obtain acute mushroom poisoning death risk related indexes; performing multi-factor Logistic regression analysis based on the acute mushroom poisoning death risk related indexes to obtain acute mushroom poisoning death independent risk factors; taking the independent risk factors of the acute mushroom poisoning death as predictive factors, and establishing a line diagram predictive model of the acute mushroom poisoning death; and carrying out distinction degree evaluation, calibration degree evaluation, decision curve evaluation and rationality analysis on the constructed column graph prediction model based on the verification set, and optimizing the column graph prediction model based on an analysis result to obtain a final column graph prediction model. The method can guide the standardized early recognition of the acute lethal mushroom poisoning process and treatment strategy, and improves the prognosis of the acute mushroom poisoning patient.

Description

Technical Field

[0001] The present invention relates to the technical field of mushroom poisoning prediction, in particular to a method for constructing an acute mushroom poisoning death prediction model. Background Art

[0002] Fatal mushroom poisoning is characterized by rapid disease progression, complex mechanisms, and high mortality rates, necessitating early identification, diagnosis, and treatment. The clinical characteristics of the liver-damaging form of fatal mushroom poisoning are divided into four phases: the incubation period, the gastroenteritis phase, the pseudo-healing phase, and the visceral damage phase. The incubation period is 6-12 hours, with symptoms occasionally appearing after 20 hours. Patients often present with gastrointestinal symptoms, and severe cases gradually develop multiple organ dysfunction. Due to the incubation period and pseudo-healing phase, liver function tests are often normal during this period. Initial physicians often fail to collect a complete history of mushroom poisoning, lack awareness of mushroom identification, and fail to promptly collect mushroom specimens for morphological and molecular identification. This can lead to misclassification of the liver-damaging form of fatal mushroom poisoning as a common gastrointestinal form. Treatment is also substandard, and the comprehensive diagnosis and treatment of acute poisoning patients varies widely across regions, potentially missing the optimal time for diagnosis and treatment. Early identification of the symptoms, clinical diagnosis, and proactive and effective treatment are crucial for improving the prognosis of patients with fatal mushroom poisoning. Although existing clinical classifications provide some guidance for patient prognosis, they are not suitable for early diagnosis and severity assessment of fatal mushroom poisoning. Therefore, it is necessary to establish a comprehensive and objective evaluation system for predicting mortality from acute mushroom poisoning. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for constructing a mortality prediction model for acute mushroom poisoning, establish a multi-factor joint prediction model through clinical data and scores, evaluate the severity of the disease and the risk of death in patients with acute mushroom poisoning, guide doctors to standardize the diagnosis and treatment process of acute mushroom poisoning, and carry out clustered treatment for fatal mushroom poisoning as soon as possible, which can effectively reduce the mortality rate of fatal mushroom poisoning. The prediction model can also avoid overtreatment of mild patients, and has good social and economic benefits.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A method for constructing a prediction model for acute mushroom poisoning mortality, comprising:

[0006] Collect test data from several people with acute mushroom poisoning and construct training and validation sets;

[0007] Performing univariate analysis based on the data of acute mushroom poisoning patients in the training set to obtain indicators related to the risk of death from acute mushroom poisoning;

[0008] A multivariate logistic regression analysis was performed based on the risk-related indicators of death from acute mushroom poisoning to obtain independent risk factors for death from acute mushroom poisoning;

[0009] Using the independent risk factors for death from acute mushroom poisoning as predictors, a nomogram prediction model for death from acute mushroom poisoning was established;

[0010] Based on the validation set, the constructed nomogram prediction model is evaluated for discrimination, calibration, decision curve, and rationality, and the nomogram prediction model is optimized based on the analysis results to obtain the final nomogram prediction model.

[0011] Optionally, the test data include clinical indicators, laboratory indicators, evaluation of HOPE6-TALK scores, evaluation of damaged organs, evaluation of severity, and prognosis and follow-up related data.

[0012] Optionally, constructing a training set and a validation set includes:

[0013] Collect test data from several people with acute mushroom poisoning;

[0014] Based on the detection data, several acute mushroom poisoning patients were screened according to preset criteria, and a data set was constructed and divided into a training set and a validation set, wherein the training set included a recovery group and a death group.

[0015] Optionally, the univariate analysis is performed based on the acute mushroom poisoning detection data in the training set to obtain indicators related to the risk of death from acute mushroom poisoning, including:

[0016] Based on the detection data of the recovered acute mushroom poisoning patients in the recovery group and the detection data of the deceased acute mushroom poisoning patients in the death group in the training set, a univariate analysis was performed to obtain indicators related to the risk of death from acute mushroom poisoning.

[0017] Optionally, the indicators related to the risk of death from acute mushroom poisoning include: age, main pathogenic factors, number of days of hospitalization, whether blood purification is performed, whether the first visit to our hospital, time from discharge symptoms to hospital, HOPE6-TALK score, potassium, AST, ALT, total bilirubin, indirect bilirubin, direct bilirubin, urea nitrogen, creatinine, creatine kinase, creatine kinase isoenzyme, lactate dehydrogenase, α-hydroxybutyric acid dehydrogenase, INR, fibrinogen, SOFA score, and APACHEII score.

[0018] Optionally, the independent risk factors for death from acute mushroom poisoning include: APACHE II score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether blood purification treatment is used.

[0019] The beneficial effects of the present invention are:

[0020] The present invention collects clinical symptoms, signs, HOPE6-TALK scores, assesses damaged organs, assesses severity, outcomes, and follow-up indicators of patients with acute mushroom poisoning. Through univariate analysis and multivariate logistic regression analysis, the independent risk factors for death from acute mushroom poisoning, such as APACHE II score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether blood purification treatment is used, are determined. A nomogram model for predicting death from acute mushroom poisoning is constructed as a predictor to assess the probability of death from acute mushroom poisoning. By constructing a concise and intuitive visual scoring system, the outcome and prognosis of acute mushroom poisoning can be judged early. Early diagnosis of acute lethal mushroom poisoning and standardized bundled treatment can effectively reduce the mortality rate and improve the prognosis. The method can guide standardized early identification of acute lethal mushroom poisoning processes and treatment strategies, improve the prognosis of patients with acute mushroom poisoning, and has important clinical application prospects and promotion value. At the same time, the prediction can also avoid overtreatment of mild patients, with good social and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a flow chart of a method for constructing an acute mushroom poisoning death prediction model according to an embodiment of the present invention;

[0023] Figure 2 Graph showing the variable screening results of the LASSO regression analysis according to an embodiment of the present invention, wherein (a) is a cross-validation curve and (b) is a LASSO coefficient curve;

[0024] Figure 3 Schematic diagram of a nomogram prediction model according to an embodiment of the present invention;

[0025] Figure 4 : ROC curve diagram of the nomogram prediction model of the embodiment of the present invention, wherein (a) is the ROC curve under the training set, and (b) is the ROC curve under the validation set;

[0026] Figure 5 Graphs showing the calibration curves of the nomogram prediction model according to an embodiment of the present invention, wherein (a) is the calibration curve under the training set, and (b) is the calibration curve under the validation set;

[0027] Figure 6: This is a DCA clinical decision curve diagram of the nomogram prediction model of an embodiment of the present invention, wherein (a) is the DCA clinical decision curve under the training set, and (b) is the DCA clinical decision curve under the validation set;

[0028] Figure 7 1 is a diagram showing the rationality analysis results of the nomogram prediction model according to an embodiment of the present invention, wherein (a) is the rationality analysis result under the training set, and (b) is the rationality analysis result under the validation set. DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0031] This embodiment provides a method for constructing a prediction model for acute mushroom poisoning death, such as Figure 1 Shown, including:

[0032] Collect test data from several people with acute mushroom poisoning and construct training and validation sets;

[0033] Performing univariate analysis based on the data of acute mushroom poisoning patients in the training set to obtain indicators related to the risk of death from acute mushroom poisoning;

[0034] A multivariate logistic regression analysis was performed based on the risk-related indicators of death from acute mushroom poisoning to obtain independent risk factors for death from acute mushroom poisoning;

[0035] Using the independent risk factors for death from acute mushroom poisoning as predictors, a nomogram prediction model for death from acute mushroom poisoning was established;

[0036] Based on the validation set, the constructed nomogram prediction model is evaluated for discrimination, calibration, decision curve, and rationality, and the nomogram prediction model is optimized based on the analysis results to obtain the final nomogram prediction model.

[0037] Specifically, this example collects clinical symptoms, signs, HOPE6-TALK scores, organ damage assessment, severity assessment, outcome, and follow-up from patients with acute mushroom poisoning. Univariate and multivariate logistic regression analyses identify independent risk factors for death from acute mushroom poisoning, including APACHE II score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether or not blood purification therapy was used. These factors are then used as predictors to construct a nomogram model for predicting death from acute mushroom poisoning. This model assesses the probability of death from acute mushroom poisoning. This concise and intuitive visual scoring system, the nomogram, allows for early assessment of the outcome and prognosis of acute mushroom poisoning, and allows for early identification of acute lethal mushroom poisoning and standardized bundled treatment to determine whether it can effectively reduce mortality and improve prognosis. This model can guide standardized early identification of acute lethal mushroom poisoning processes and treatment strategies, improve the prognosis of patients with acute mushroom poisoning, and has significant clinical application prospects and promotional value. Furthermore, this model can also avoid overtreatment of mild patients through prediction, resulting in significant social and economic benefits.

[0038] Furthermore, constructing the training set and validation set includes:

[0039] Collect test data from several patients with acute mushroom poisoning, including clinical indicators, laboratory indicators, HOPE6-TALK score assessment, assessment of damaged organs, assessment of severity, and outcome and follow-up data;

[0040] Based on the detection data, several acute mushroom poisoning patients were screened according to preset criteria, and a data set was constructed and divided into a training set and a validation set, wherein the training set included a recovery group and a death group.

[0041] Furthermore, based on the univariate analysis of the acute mushroom poisoning detection data in the training set, the indicators related to the risk of death from acute mushroom poisoning were obtained, including:

[0042] Based on the detection data of the recovered acute mushroom poisoning patients in the recovery group and the detection data of the deceased acute mushroom poisoning patients in the death group in the training set, a univariate analysis was performed to obtain indicators related to the risk of death from acute mushroom poisoning.

[0043] Among them, the indicators related to the risk of death from acute mushroom poisoning include: age, main pathogenic factors, number of days of hospitalization, whether blood purification is performed, whether the patient is the first visit to our hospital, time from discharge symptoms to hospital, HOPE6-TALK score, potassium, AST, ALT, total bilirubin, indirect bilirubin, direct bilirubin, urea nitrogen, creatinine, creatine kinase, creatine kinase isoenzyme, lactate dehydrogenase, α-hydroxybutyric acid dehydrogenase, INR, fibrinogen, SOFA score, and APACHEII score.

[0044] Furthermore, the independent risk factors for death from acute mushroom poisoning include: APACHE II score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether blood purification treatment was performed.

[0045] Specifically, the prediction model construction process of this embodiment is as follows: Figure 1 Shown, including:

[0046] Step 1: Obtain clinical and laboratory indicators, HOPE6-TALK score, damaged organ assessment, severity assessment, outcome, follow-up and other relevant data of patients with mushroom poisoning. Randomly divide the data into 70% as training set and 30% as validation set for baseline analysis.

[0047] Specifically, the detection data of acute mushroom poisoning patients at Zunyi Medical University from July 2013 to December 2022 were collected. The data were randomly divided into training set and validation set in a ratio of 7:3 using R 4.4.2 software, and the baseline comparability was verified. The training set included the recovery group and the death group.

[0048] Step 2: Perform univariate analysis on the relevant data in the training set to screen and obtain indicators related to mushroom poisoning deaths.

[0049] Univariate analysis was specifically a difference analysis. By comparing the basic information of the recovery and death groups of acute mushroom poisoning patients, p ≤ 0.05 was used to screen out mushroom poisoning death-related indicators such as age, main pathogenic factors, length of hospital stay, whether blood purification was performed, whether the first visit was in our hospital, time from discharge symptoms to hospital, HOPE6 / TALK score, potassium, AST, ALT, total bilirubin, indirect bilirubin, direct bilirubin, urea nitrogen, creatinine, creatine kinase, creatine kinase isoenzyme, lactate dehydrogenase, α-hydroxybutyric acid dehydrogenase, INR, fibrinogen, SOFA score, and APACHE II score.

[0050] Step 3, the variable screening results of LASSO regression analysis are as follows Figure 2 As shown in the figure, (a) is the cross-validation curve, and (b) is the LASSO coefficient curve. To eliminate the possibility of collinearity among the variables during data analysis and to include indicators with differences between groups, the R language glmnet package was used to fit the lasso regression. By introducing a penalty term, the variable coefficients were continuously compressed to remove redundant variables. The λ value was calculated according to the cross-validation method, and the optimal λ was used as the criterion to screen characteristic variables. Based on the above-screened characteristic variables, multivariate logistic regression analysis was performed to obtain seven independent risk factors for death from acute mushroom poisoning, including APACHE II score, SOFA score, HOPE6-TALK score, length of hospital stay, whether blood purification treatment was used, AST, and creatine kinase.

[0051] A multivariate logistic regression analysis was established for the factors related to death from acute mushroom poisoning. APACHEII score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether blood purification treatment was used were independent risk factors for death from acute mushroom poisoning. Using the above indicators as predictors, a nomogram prediction model for death from acute mushroom poisoning was established, such as Figure 3 shown.

[0052] Application of the nomogram prediction model: When a patient poisoned by eating mushrooms seeks medical treatment, the patient's APACHE II score, SOFA score, HOPE6-TALK score, whether blood purification treatment was performed, number of days in hospital, AST, and creatine kinase test values can be vertically aligned with the top score line according to the nomogram. The values of the seven variables can be assigned as P1, P2,...P7 respectively. The total score is calculated by adding up the values of the seven variables. The vertical line aligned with the Diagnostic Possibility scale is the patient's probability of survival.

[0053] Step 4: Perform discrimination evaluation on the nomogram prediction model constructed above to test the model's ability to predict the probability of death from acute mushroom poisoning.

[0054] To evaluate the performance of the model, the ROC curve was drawn based on the nomogram prediction model. Figure 4 As shown in the figure, (a) is the ROC curve for the training set, and (b) is the ROC curve for the validation set. The area under the curve (AUC) was calculated to evaluate the model's performance. The results showed that in the training set, the AUC and its 95% confidence interval were 0.980 (0.960-1.000); in the validation set, the AUC and its 95% confidence interval were 0.979 (0.959-1.000), indicating that the model has high predictive accuracy.

[0055] Step 5: Conduct calibration evaluation, decision curve assessment, and rationality analysis on the nomogram prediction model constructed above.

[0056] The calibration of the prediction model is evaluated using the calibration curve as an indicator, such as Figure 5 As shown in the figure, (a) is the calibration curve under the training set, and (b) is the calibration curve under the validation set. In the training set and validation set, the calibration curve fits well with the ideal curve, and the predicted probability and observed probability are similar, indicating that the model has good fit and good predictive ability. In order to evaluate the clinical practicality of the model, the DCA decision curve is used as the evaluation indicator of the prediction model, as shown in the figure. Figure 6As shown, (a) is the DCA clinical decision curve under the training set, and (b) is the DCA clinical decision curve under the validation set. The horizontal axis represents the threshold probability and the vertical axis represents the net benefit. Compared with no intervention and full intervention, the DCA curves in the training set and validation set show that when the threshold probability is within the range of 10% to 100%, patients can benefit from this prediction model, indicating that the model has good clinical practical value. Finally, the ROC curves of this prediction model and the seven risk factors screened by multivariate logistic regression are drawn respectively, as shown below. Figure 7 As shown in the figure, (a) is the rationality analysis result under the training set, and (b) is the rationality analysis result under the validation set. In both the training set and the validation set, the area under the curve of this model is larger than the area under the curve of any risk factor, indicating that this model has better risk prediction ability.

[0057] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A method for constructing a prediction model for acute mushroom poisoning mortality, characterized in that: include: Collect test data from several people with acute mushroom poisoning and construct training and validation sets; Performing univariate analysis based on the data of acute mushroom poisoning patients in the training set to obtain indicators related to the risk of death from acute mushroom poisoning; A multivariate logistic regression analysis was performed based on the risk-related indicators of death from acute mushroom poisoning to obtain independent risk factors for death from acute mushroom poisoning; Using the independent risk factors for death from acute mushroom poisoning as predictors, a nomogram prediction model for death from acute mushroom poisoning was established; Based on the validation set, the constructed nomogram prediction model is evaluated for discrimination, calibration, decision curve, and rationality, and the nomogram prediction model is optimized based on the analysis results to obtain the final nomogram prediction model.

2. The method for constructing a prediction model for acute mushroom poisoning death according to claim 1, characterized in that: The test data include clinical indicators, laboratory indicators, evaluation of HOPE6-TALK score, evaluation of damaged organs, evaluation of severity, as well as outcome and follow-up related data.

3. The method for constructing a prediction model for acute mushroom poisoning death according to claim 1, characterized in that: Constructing the training set and validation set includes: Collect test data from several people with acute mushroom poisoning; Based on the detection data, several acute mushroom poisoning patients were screened according to preset criteria, and a data set was constructed and divided into a training set and a validation set, wherein the training set included a recovery group and a death group.

4. The method for constructing a prediction model for acute mushroom poisoning death according to claim 3, characterized in that: Based on the univariate analysis of the acute mushroom poisoning test data in the training set, the indicators related to the risk of death from acute mushroom poisoning were obtained, including: Based on the detection data of the recovered acute mushroom poisoning patients in the recovery group and the detection data of the deceased acute mushroom poisoning patients in the death group in the training set, a univariate analysis was performed to obtain indicators related to the risk of death from acute mushroom poisoning.

5. The method for constructing a prediction model for acute mushroom poisoning death according to claim 4, characterized in that: The risk-related indicators of death from acute mushroom poisoning include: age, main pathogenic factors, number of days of hospitalization, whether blood purification is performed, whether the first visit to our hospital, time from discharge symptoms to hospitalization, HOPE6-TALK score, potassium, AST, ALT, total bilirubin, indirect bilirubin, direct bilirubin, urea nitrogen, creatinine, creatine kinase, creatine kinase isoenzyme, lactate dehydrogenase, α-hydroxybutyric acid dehydrogenase, INR, fibrinogen, SOFA score, and APACHEII score.

6. The method for constructing a prediction model for acute mushroom poisoning death according to claim 1, characterized in that: The independent risk factors for death from acute mushroom poisoning include: APACHE II score, SOFA score, length of hospital stay, AST, creatine kinase, HOPE6-TALK score, and whether blood purification treatment was performed.

Images