CN-121983322-A - Liver and gall postoperative infection risk grading early warning method based on clinical parameters

CN121983322ACN 121983322 ACN121983322 ACN 121983322ACN-121983322-A

Abstract

The invention relates to the field of liver and gall postoperative infection risk grading early warning, in particular to a liver and gall postoperative infection risk grading early warning method based on clinical parameters. The method comprises the steps of collecting original clinical parameters and preprocessing the same to obtain preprocessed standardized clinical parameters, introducing clinical basic weights, applying a multidimensional distribution mapping algorithm based on the preprocessed standardized clinical parameters to calculate an infection risk score, classifying risk grades based on the infection risk score, introducing a dynamic weighting mechanism for emphasizing effective updating, and dynamically updating the clinical basic weights. The method solves the problems that the traditional liver and gall postoperative infection risk assessment method ignores complex high-order interaction and nonlinear association among indexes such as body temperature, white blood cell count, procalcitonin, C-reactive protein and the like, parameter weights cannot be adjusted in real time according to the dynamic change of postoperative illness state of a patient, and definite risk classification rules are lacked, so that engineering application value is limited.

Inventors

LI SHUANG
YANG ZHAOXU

Assignees

中国人民解放军空军军医大学

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (9)

1. The liver and gall postoperative infection risk grading early warning method based on clinical parameters is characterized by comprising the following steps of: S1, acquiring original clinical parameters and preprocessing to obtain preprocessed standardized clinical parameters, introducing a multidimensional distribution mapping algorithm, quantifying abnormal deviation degree of the preprocessed standardized clinical parameters based on the preprocessed standardized clinical parameters, introducing clinical basic weight, weighting the abnormal deviation degree of the preprocessed standardized clinical parameters to obtain weighted abnormal deviation degree, correcting the weighted abnormal deviation degree in time to generate a grading magnitude, and calculating an infection risk grade; S2, dividing risk grades based on infection risk scores and obtaining an infection result, introducing a dynamic weighting mechanism for emphasizing effective updating, constructing a significance threshold item based on the pretreated standardized clinical parameters and combining a reference mean value of the pretreated standardized clinical parameters, and combining a hyperbolic tangent function to generate a weight adjustment factor of the dynamic weighting mechanism, dynamically updating clinical basic weights based on the weight adjustment factor of the dynamic weighting mechanism, and balancing the stability and sensitivity of infection risk score calculation.
2. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S1 specifically comprises the following steps: The pretreatment process comprises missing value filling, outlier filtering and standardization treatment.
3. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S1 specifically comprises the following steps: The method comprises the steps of introducing postoperative standardized clinical parameters of historical liver and gall postoperative infection cases, quantifying objective weights of all clinical parameters on infection risk assessment through an entropy weight method, obtaining subjective importance weights of all clinical parameters through a hierarchical analysis method, and finally carrying out coupling optimization on the objective weights and the subjective importance weights through a weighted summation method to obtain clinical basic weights.
4. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S1 specifically comprises the following steps: in the implementation process of the multidimensional distribution mapping algorithm, based on the preprocessed standardized clinical parameters, the abnormal deviation degree of the preprocessed standardized clinical parameters is quantized by combining Gaussian distribution.
5. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S1 specifically comprises the following steps: In the implementation process of the multidimensional distribution mapping algorithm, a time correction term is introduced to adjust the weighted abnormal deviation degree, and the time correction term is obtained by introducing a time sensitivity coefficient and combining the ratio of the monitored duration after operation to the total observation period after operation.
6. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 5, wherein the step S1 specifically comprises: The infection risk score is calculated as follows: ; wherein R (t) is an infection risk score at a time point t, n is the total number of types of the pretreated standardized clinical parameters actually required to be observed in clinical regulation, i is the type index of the pretreated standardized clinical parameters; Clinical basis weights for standardized clinical parameters after class i pretreatment at time point t; actual values of standardized clinical parameters after class i pretreatment at time point t; Is the reference mean value of the standardized clinical parameters after the i-th pretreatment, The standard deviation of the standardized clinical parameters after the i-th pretreatment is used as the reference standard deviation; engineering amplification factors of standardized clinical parameters after the i-th pretreatment; Is a time sensitivity coefficient; the post-operation monitored duration is the current time point t; Total observation period after operation; abnormal deviation degree of the standardized clinical parameters after pretreatment; Is a time correction term; Representing the magnitude of the score.
7. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S2 specifically comprises the following steps: Based on the infection risk score and the threshold value of the infection risk grade division, the postoperative infection risk of the patient is divided into three grades of low, medium and high, the infection results corresponding to the medium risk grade and the high risk grade are uniformly regarded as infection, and the infection results corresponding to the low risk grade are regarded as non-infection.
8. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 1, wherein the step S2 specifically comprises the following steps: in the implementation process of the dynamic weighting mechanism for emphasizing effective update, a weight adjustment coefficient is introduced, and a hyperbolic tangent function is applied to the significance threshold item to generate a weight adjustment factor of the dynamic weighting mechanism.
9. The method for grading and early warning of risk of infection after liver and gall surgery based on clinical parameters according to claim 8, wherein the step S2 specifically comprises: The specific calculation mode of the weight adjustment coefficient comprises the steps of setting an initial candidate interval of the weight adjustment coefficient, dividing postoperative standardized clinical parameters of historical hepatobiliary postoperative infection cases into independent data subsets, aiming at each candidate value in the candidate interval, testing stability and sensitivity of a calculation formula of infection risk scores on different data subsets, screening to obtain the optimal weight adjustment coefficient, wherein the stability takes oscillation amplitude of clinical basic weight as an evaluation index, sensitivity takes risk prediction accuracy as an evaluation index, comparing an infection result of the calculated cases with a real infection result of a corresponding case in postoperative standardized clinical parameters of the historical hepatobiliary postoperative infection cases, obtaining the number of cases with the same infection result as the real infection result, and obtaining the risk prediction accuracy based on the ratio of the number of cases with the same infection result as the real infection result to the total number of cases participating in the test.

Description

Liver and gall postoperative infection risk grading early warning method based on clinical parameters Technical Field The invention relates to the field of liver and gall postoperative infection risk grading early warning, in particular to a liver and gall postoperative infection risk grading early warning method based on clinical parameters. Background The liver and gall surgery belongs to the field of clinical high-risk surgery due to the characteristics of special operation parts, high operation complexity, various basic conditions of patients and the like, and postoperative infection such as incision infection, abdominal infection, biliary tract infection and the like is a common complication. The occurrence of postoperative infection not only can prolong the hospitalization period of patients and increase the medical cost burden, but also can cause severe secondary diseases such as infectious shock, multi-organ failure and the like, and can cause severe influence on prognosis and medical quality of the patients. Therefore, the method can evaluate the infection risk of the patients after the hepatobiliary surgery timely and accurately, provide scientific basis for clinical intervention, and is one of the core demands of hepatobiliary surgery clinical monitoring and diagnosis and treatment work. At present, a plurality of methods for evaluating the risk of infection after liver and gall surgery exist in the clinical and scientific research fields, the core of the methods is unfolded around clinical parameters such as postoperative physiological indexes, inflammatory indexes and the like, the risk quantification and the prognosis are realized by means of various modeling methods, and a certain support is provided for postoperative infection prevention and control. However, most of traditional methods for evaluating risk of infection after liver and gall surgery rely on traditional statistical models such as linear regression and logistic regression, and the core limitation is that only simple linear relations among clinical parameters can be captured, and complex high-order interactions and nonlinear correlations among indexes such as body temperature, white blood cell count, procalcitonin, C-reactive protein and the like are ignored. Meanwhile, the parameter weight of the traditional method is a fixed value, and the traditional method cannot be adjusted in real time according to the dynamic change of the postoperative condition of a patient, so that the model has poor adaptability and insufficient prediction precision, and the requirement of clinical real-time monitoring is difficult to meet. In addition, part of the assessment methods lack clear risk classification rules, output results are difficult to directly guide the formulation of clinical intervention measures, and engineering application values are limited. Disclosure of Invention The invention provides a liver and gall postoperative infection risk grading early warning method based on clinical parameters, which aims to solve the technical problems that the traditional liver and gall postoperative infection risk assessment method can only capture simple linear relations among clinical parameters, ignores complex high-order interactions and nonlinear correlations among indexes such as body temperature, leucocyte count, procalcitonin, C-reactive protein and the like, has a parameter weight of a fixed value, cannot be adjusted in real time according to the dynamic change of postoperative illness of a patient, lacks clear risk grading rules, and has difficult direct guidance on the formulation of clinical intervention measures and limited engineering application value. The invention discloses a liver and gall postoperative infection risk grading early warning method based on clinical parameters, which comprises the following steps of: S1, acquiring original clinical parameters and preprocessing to obtain preprocessed standardized clinical parameters, introducing a multidimensional distribution mapping algorithm, quantifying abnormal deviation degree of the preprocessed standardized clinical parameters based on the preprocessed standardized clinical parameters, introducing clinical basic weight, weighting the abnormal deviation degree of the preprocessed standardized clinical parameters to obtain weighted abnormal deviation degree, correcting the weighted abnormal deviation degree in time to generate a grading magnitude, and calculating an infection risk grade; S2, dividing risk grades based on infection risk scores and obtaining an infection result, introducing a dynamic weighting mechanism for emphasizing effective updating, constructing a significance threshold item based on the pretreated standardized clinical parameters and combining a reference mean value of the pretreated standardized clinical parameters, and combining a hyperbolic tangent function to generate a weight adjustment factor of the dynamic weighting mechanism, dynamically updating