Dynamic prediction algorithm for monitoring late-onset infection of premature infant and upgrading system

Abstract

The invention provides a dynamic prediction algorithm for monitoring late-onset infection of a premature infant and an upgrading system. The dynamic prediction algorithm for monitoring late-onset infection of the premature infant and the upgrading system comprise the following steps: a, collecting clinical data of the premature infant, including birth weight, gestational age, 1-minute and 5-minute Apgar scores, right hand perfusion index, lower limb perfusion index and other related clinical information, and b, determining the early-onset infection of the premature infant through medical history collection and physical sign analysis. The method comprises the following steps: collecting data of 11 classification independent variables: prenatal antibiotic use conditions (existence and absence); according to the dynamic prediction algorithm for monitoring late-onset infection of the premature infant and the upgrading system, the infection risk index is effectively calculated through high-risk factors analyzed by the Lasso regression model in combination with clinical basic data of the premature infant, early warning of infection of the premature infant is provided for medical staff, and the accuracy and timeliness of infection prediction are remarkably improved. Besides, the system can automatically remind medical staff to intervene the high-risk child patient through red warning, so that the death rate caused by delayed discovery and delayed treatment is reduced, and the clinical intervention effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic prediction algorithms for late-onset infections, and in particular to a dynamic prediction algorithm and an upgrade system for monitoring late-onset infections in premature infants. Background Art

[0002] Late-onset sepsis (LOS) refers to sepsis that occurs in newborns 72 hours after birth and is one of the main causes of neonatal morbidity and mortality. Especially for very low birthweight infants (VLBWI), due to their incomplete immune system and organ development, the incidence of sepsis is as high as 20% to 54%, and the mortality rate is 18%. In the case of concurrent septic shock, the mortality rate can reach 60%. Even surviving infants often face the risk of short-term complications and long-term developmental disorders, including impaired neurodevelopment, intellectual and psychomotor developmental delays, etc. Studies have shown that LOS is clearly associated with neurodevelopmental disorders such as cerebral palsy and visual impairment that appear in VLBWI at 18 months of age.

[0003] Because the early symptoms of neonatal sepsis lack specificity and clinical signs are often subtle, early identification becomes difficult, especially in very low birth weight infants. Although studies have shown that factors such as gestational age, birth weight, central venous catheterization, and endotracheal intubation are closely related to the development of LOS in very low birth weight infants, there is currently a lack of effective clinical indicators that can reliably predict early infection. Some children unfortunately die due to delayed initiation of antibiotic treatment. For newborns at risk of infection, if treatment is started within 1 hour, the risk of death can be reduced by 10%, and if treatment is started within 6 hours, the survival rate of the child can be increased by approximately 30%. Therefore, early prediction of the risk of LOS in very low birth weight infants is of great significance for clinicians to timely adjust the timing of antibiotic treatment, reduce mortality, and improve the prognosis of surviving infants. Summary of the Invention

[0004] In view of the current lack of automatic early warning monitoring technology for premature infant infections in clinical practice, the present invention provides a dynamic prediction algorithm and upgrade system for monitoring late-onset infections in premature infants, which solves the problem of lack of effective indicators that can reliably predict early infections in clinical practice and the unfortunate death of some children due to delayed initiation of antibiotic treatment.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic prediction algorithm and upgrade system for monitoring late-onset infections in premature infants, comprising the following steps:

[0006] a. Collect clinical data of premature infants, including birth weight, gestational age, 1-minute and 5-minute Apgar scores, right hand perfusion index, lower limb perfusion index, and other relevant clinical information;

[0007] b. Data on the following 11 categorical independent variables were collected through history collection and physical sign analysis: antenatal antibiotic use (1 = yes, 0 = no), umbilical arteriovenous catheterization (1 = yes, 0 = no), endotracheal intubation (1 = yes, 0 = no), small for gestational age (1 = yes, 0 = no), poor responsiveness (1 = yes, 0 = no), skin color change (1 = yes, 0 = no), abdominal distension (1 = yes, 0 = no), abnormal blood sugar (1 = yes, 0 = no), abnormal white blood cell count (1 = yes, 0 = no), elevated procalcitonin (1 = yes, 0 = no), and elevated C-reactive protein (1 = yes, 0 = no).

[0008] c. The above 17 independent variables were input into the Lasso regression analysis model to calculate the regression coefficients and obtain the infection risk index for late-onset sepsis in neonates. The regression model formula is:

[0009]

[0010] in:

[0011] y is the infection risk index of late-onset sepsis in neonates;

[0012] β0 is the intercept term;

[0013] β i is the regression coefficient corresponding to the i-th independent variable, including birth weight, gestational age, Apgar score, and perfusion index;

[0014] X i Seventeen clinical independent variables were entered, including birth weight, gestational age, Apgar score, and perfusion index;

[0015] d. Conduct dynamic monitoring based on the calculated infection risk index and determine whether to take warning measures based on the set threshold.

[0016] Preferably, the algorithm comprises the following steps:

[0017] f. Based on the real-time updated risk index, the system can automatically analyze and determine the risk level. If the risk index exceeds the preset threshold, the system triggers a red alert, reminding clinical medical staff to intervene further;

[0018] g. When the risk index is updated, the system can automatically adjust the weight of the risk prediction model based on the patient's clinical condition, continuously optimize the prediction effect, and ensure the accuracy of early predictions;

[0019] h. The system has real-time data collection and processing capabilities, can handle data input from different devices, and promptly update risk prediction information based on different data sources;

[0020] i. After a red alert is triggered, the system automatically generates a report and displays it through the interface. The report includes the patient's clinical data, risk analysis, and recommended treatment options to assist in clinical decision-making;

[0021] j. The system also has historical data recording and analysis functions, which can conduct retrospective analysis of past data and provide trend forecasts for clinicians to help monitor changes in patients' health status.

[0022] Preferably, the Lasso regression analysis model constrains the regression coefficients through L1 regularization, so that the coefficients of unimportant variables tend to zero, thereby achieving variable selection and model optimization.

[0023] A system for implementing the algorithm, comprising:

[0024] Data collection module, used to automatically collect birth weight, gestational age, 1-minute and 5-minute Apgar scores, perfusion index, categorical independent variables and other relevant clinical data of premature infants;

[0025] The risk calculation module is used to input the 17 collected independent variables into the Lasso regression analysis model to calculate the regression coefficient and obtain the infection risk index of neonatal late-onset sepsis;

[0026] The alarm module automatically issues a red alert when the calculated infection risk index exceeds the set threshold, reminding clinical medical staff that the newborn may be at risk of infection;

[0027] A display module is used to display the calculated risk index in real time and update the display content based on clinical data and risk index;

[0028] The storage module is used to store all collected clinical data, risk calculation results and historical data, and can generate reports or trend analysis for subsequent evaluation and decision-making.

[0029] Preferably, it further comprises a real-time updating module for updating the input independent variable data according to the clinical status of the patient and recalculating the infection risk index;

[0030] Data processing module, used to clean and normalize clinical data collected from different data sources to ensure data quality and consistency;

[0031] Model optimization module, used to adjust the parameters of the Lasso regression model based on historical data and clinical feedback to improve the accuracy of prediction;

[0032] The risk trend analysis module is used to display the infection risk trend of each child at different time points, assisting doctors in judging changes in the child's condition.

[0033] Preferably, the alarm module further includes:

[0034] Multi-level alert function: when the infection risk index reaches different thresholds, the system can issue yellow, orange and red alerts in sequence, and the alert level is graded according to the severity of the risk;

[0035] Alarm response function: after the system issues an alarm, it can guide clinical medical staff to the patient information interface and provide detailed diagnostic suggestions and treatment plans.

[0036] Preferably, the formula for the L1 regularization term of the Lasso regression model is:

[0037]

[0038] The regularization term penalizes excessively large regression coefficients, prompting the model to select a small number of independent variables with significant influence, thereby avoiding overfitting and improving the interpretability and predictive ability of the model.

[0039] Preferably, the Lasso regression model can automatically adjust the regression coefficient according to the clinical data input in real time, optimize the prediction effect and reduce the interference of unimportant variables.

[0040] The present invention provides a dynamic prediction algorithm and upgrade system for monitoring late-onset infections in premature infants. It has the following beneficial effects:

[0041] This dynamic prediction algorithm and upgraded system for monitoring late-onset infections in premature infants uses a Lasso regression model, combined with clinical data from premature infants, to effectively calculate an infection risk index, providing doctors with early warnings and significantly improving the accuracy and timeliness of infection predictions. Furthermore, the system automatically alerts medical staff to intervene in high-risk infants through red alerts, thereby reducing mortality due to delayed treatment and improving the effectiveness of clinical interventions.

[0042] The system of the present invention also has powerful real-time data processing and optimization functions, and continuously optimizes the prediction effect through dynamically updated risk prediction models. The system can automatically collect, process and update data based on different data sources, such as the hospital's electronic health records and monitoring equipment, to ensure the timeliness and accuracy of clinical data. Through the L1 regularized Lasso regression model, the algorithm can adaptively select important variables and reduce the influence of irrelevant variables, effectively avoid overfitting, and improve the interpretability and predictive ability of the model. The multi-level alarm function ensures that medical staff can take quick action according to different risk levels, while the trend analysis module helps doctors monitor the health changes of children in real time and provide more accurate treatment plans. Therefore, this technical solution not only enhances the reliability of early infection prediction, but also effectively improves the intelligence and accuracy of clinical decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a structural schematic diagram of the present invention.

[0044] Figure 2 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0045] 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] Example 1

[0047] like Figure 1-Figure 2 As shown, the embodiment of the present invention provides a dynamic prediction algorithm and upgrade system for monitoring late-onset infections in premature infants, including the following steps:

[0048] a. Collect clinical data of premature infants, including birth weight, gestational age, 1-minute and 5-minute Apgar scores, right hand perfusion index, lower limb perfusion index, and other relevant clinical information.

[0049] b. Data on the following 11 categorical independent variables were collected through history collection and physical sign analysis: antenatal antibiotic use (1 = yes, 0 = no), umbilical arteriovenous catheterization (1 = yes, 0 = no), endotracheal intubation (1 = yes, 0 = no), small for gestational age (1 = yes, 0 = no), poor responsiveness (1 = yes, 0 = no), skin color change (1 = yes, 0 = no), abdominal distension (1 = yes, 0 = no), abnormal blood sugar (1 = yes, 0 = no), abnormal white blood cell count (1 = yes, 0 = no), elevated PCT (1 = yes, 0 = no), and elevated CRP (1 = yes, 0 = no).

[0050] c. The above 17 independent variables were input into the Lasso regression analysis model to calculate the regression coefficients and obtain the infection risk index for late-onset sepsis in neonates. The regression model formula is:

[0051]

[0052] in:

[0053] y is the infection risk index of late-onset sepsis in neonates.

[0054] β0 is the intercept term.

[0055] β i is the regression coefficient corresponding to the i-th independent variable, including birth weight, gestational age, Apgar score, and perfusion index.

[0056] X i For the 17 clinical independent variables input, including birth weight, gestational age, Apgar score, and perfusion index, the Lasso regression analysis model constrained the regression coefficients through L1 regularization, so that the coefficients of unimportant variables tended to zero, thereby achieving variable selection and model optimization.

[0057] d. Dynamically monitor the infection risk index and determine whether to take warning measures based on the set threshold. The algorithm includes the following steps:

[0058] f. Based on the real-time updated risk index, the system can automatically analyze and determine the risk level. If the risk index exceeds the preset threshold, the system triggers a red alert to remind clinical medical staff to make further interventions.

[0059] g. When the risk index is updated, the system can automatically adjust the weight of the risk prediction model according to the patient's clinical condition, continuously optimize the prediction effect, and ensure the accuracy of early predictions.

[0060] h. The system has real-time data collection and processing capabilities, can handle data input from different devices, and timely update risk prediction information based on different data sources.

[0061] i. After a red alert is triggered, the system automatically generates a report and displays it through the interface. The report includes the patient's clinical data, risk analysis, and recommended treatment options to assist in clinical decision-making.

[0062] j. The system also has historical data recording and analysis functions, which can conduct retrospective analysis of past data and provide trend forecasts for clinicians to help monitor changes in patients' health status.

[0063] A system for implementing an algorithm, comprising:

[0064] The data collection module is used to automatically collect the birth weight, gestational age, 1-minute and 5-minute Apgar scores, perfusion index, categorical independent variables and other relevant clinical data of premature infants.

[0065] The risk calculation module is used to input the 17 collected independent variables into the Lasso regression analysis model to calculate the regression coefficient and obtain the infection risk index of neonatal late-onset sepsis.

[0066] The alarm module automatically issues a red alert when the calculated infection risk index exceeds a set threshold, alerting clinical staff that the newborn may be at risk of infection. The alarm module also includes:

[0067] Multi-level alarm function: when the infection risk index reaches different thresholds, the system can issue yellow, orange and red warnings in turn, and the warning level is graded according to the severity of the risk.

[0068] Alarm response function: after the system issues an alarm, it can guide clinical medical staff to the patient information interface and provide detailed diagnostic suggestions and treatment plans.

[0069] The display module is used to display the calculated risk index in real time and update the display content according to the clinical data and risk index.

[0070] The storage module is used to store all collected clinical data, risk calculation results and historical data, and can generate reports or trend analysis for subsequent evaluation and decision-making.

[0071] The real-time updating module is used to update the input independent variable data according to the patient's clinical status and recalculate the infection risk index.

[0072] The data processing module is used to clean and normalize clinical data collected from different data sources to ensure data quality and consistency.

[0073] The model optimization module is used to adjust the parameters of the Lasso regression model based on historical data and clinical feedback to improve the accuracy of prediction.

[0074] The risk trend analysis module is used to display the infection risk trend of each child at different time points, assisting doctors in judging changes in the child's condition.

[0075] The formula for the L1 regularization term of the Lasso regression model is:

[0076]

[0077] The regularization term penalizes excessively large regression coefficients, prompting the model to select a small number of significantly influential independent variables, thereby avoiding overfitting and improving the model's interpretability and predictive ability. The Lasso regression model can automatically adjust the regression coefficient based on real-time clinical data input, optimize the prediction effect and reduce the interference of unimportant variables.

[0078] Example 2: Experimental Example based on this technical solution

[0079] Purpose of the experiment:

[0080] Verify the effectiveness of the dynamic prediction algorithm and upgraded system for monitoring late-onset infections in premature infants proposed in this invention in practical applications, including accuracy, real-time performance, and effectiveness of clinical intervention.

[0081] Experimental design:

[0082] This experiment selected a group of neonatal patients, including premature infants and very low birth weight infants (VLBWI), all admitted to the hospital's neonatal department. By monitoring and analyzing the clinical data of these infants, the effectiveness of the proposed algorithm in actual clinical application was evaluated.

[0083] Experimental steps:

[0084] Data collection:

[0085] After the patient is admitted to the hospital, the system automatically collects clinical data of each premature baby, including birth weight, gestational age, 1-minute and 5-minute Apgar scores, right hand perfusion index, lower limb perfusion index, etc.

[0086] At the same time, by collecting medical history and physical signs, the system obtained data on 11 categorical independent variables related to infection, including the use of antenatal antibiotics, umbilical arteriovenous catheterization, endotracheal intubation, small for gestational age, poor response, skin color change, abdominal distension, abnormal blood sugar, abnormal white blood cell count, elevated PCT, elevated CRP, etc.

[0087] Risk Calculation:

[0088] The 17 collected clinical independent variables were input into the Lasso regression analysis model to calculate the regression coefficients.

[0089] L1 regularization is used to constrain the regression coefficients, eliminate unimportant variables, optimize the model, and ensure the efficiency and accuracy of the results.

[0090] The infection risk index of each child is obtained and displayed dynamically in the system.

[0091] Real-time monitoring and alerts:

[0092] The system automatically analyzes and determines infection risks based on a real-time updated risk index. When the risk index exceeds a preset threshold, the system triggers a red alert.

[0093] After medical staff receive the alert, the system automatically generates a report that lists in detail the clinical data, risk analysis and recommended treatment plans to assist in clinical decision-making.

[0094] After a red alert is triggered, the system provides an alarm response function that can guide medical staff to the patient's detailed information interface and provide more in-depth diagnostic suggestions.

[0095] Risk trend analysis:

[0096] The system can record and display each patient's risk index trend. Through retrospective analysis of historical data, the system helps clinicians identify changes in a child's condition in real time and adjust treatment plans promptly.

[0097] Data optimization and feedback:

[0098] Each time the system updates the risk prediction, it automatically adjusts the weight of the risk prediction model based on the patient's clinical status, continuously optimizes the prediction effect, and ensures the accuracy of early predictions.

[0099] Experimental results:

[0100] accuracy:

[0101] During the experiment, the system processed 17 clinical independent variables using a Lasso regression model and successfully calculated each child's infection risk index. The experiment showed that the system's prediction accuracy reached over 85%, effectively distinguishing between high-risk and low-risk newborns.

[0102] Real-time:

[0103] The system in the experiment was able to process and display each patient's infection risk index in a short period of time (usually 1-2 minutes). The system's response speed and real-time performance were highly praised by clinical medical staff.

[0104] Warning effect:

[0105] The system successfully triggered red alerts for high-risk newborns and promptly alerted medical staff. All high-risk newborns received prompt treatment, ensuring timely treatment. The system's alert function effectively reduced the incidence of delayed antibiotic treatment.

[0106] Optimization effect:

[0107] With the continuous input and feedback of clinical data, the system can adaptively adjust the parameters of the Lasso regression model to optimize prediction accuracy. By analyzing historical data, the system can display the risk trends of each child, helping doctors to carry out long-term monitoring and intervention.

[0108] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic prediction algorithm for monitoring late-onset infections in premature infants, characterized by: The following steps are involved: a. Collect clinical data of premature infants, including birth weight, gestational age, 1-minute and 5-minute Apgar scores, right hand perfusion index, lower limb perfusion index, and other relevant clinical information; b. Data on the following 11 categorical independent variables were collected through history collection and physical sign analysis: antenatal antibiotic use (1 = yes, 0 = no), umbilical arteriovenous catheterization (1 = yes, 0 = no), endotracheal intubation (1 = yes, 0 = no), small for gestational age (1 = yes, 0 = no), poor responsiveness (1 = yes, 0 = no), skin color change (1 = yes, 0 = no), abdominal distension (1 = yes, 0 = no), abnormal blood sugar (1 = yes, 0 = no), abnormal white blood cell count (1 = yes, 0 = no), elevated procalcitonin (1 = yes, 0 = no), and elevated C-reactive protein (1 = yes, 0 = no). c. The above 17 independent variables were input into the Lasso regression analysis model to calculate the regression coefficients and obtain the infection risk index for late-onset sepsis in neonates. The regression model formula is: in: y is the infection risk index of late-onset sepsis in neonates; β0 is the intercept term; β i is the regression coefficient corresponding to the i-th independent variable, including birth weight, gestational age, Apgar score, and perfusion index; X i Seventeen clinical independent variables were entered, including birth weight, gestational age, Apgar score, and perfusion index; d. Conduct dynamic monitoring based on the calculated infection risk index and determine whether to take warning measures based on the set threshold.

2. The dynamic prediction algorithm for monitoring late-onset infections in premature infants according to claim 1, characterized in that: The algorithm comprises the following steps: f. Based on the real-time updated risk index, the system can automatically analyze and determine the risk level. If the risk index exceeds the preset threshold, the system triggers a red alert, reminding clinical medical staff to intervene further; g. When the risk index is updated, the system can automatically adjust the weight of the risk prediction model based on the patient's clinical condition, continuously optimize the prediction effect, and ensure the accuracy of early predictions; h. The system has real-time data collection and processing capabilities, can handle data input from different devices, and promptly update risk prediction information based on different data sources; i. After a red alert is triggered, the system automatically generates a report and displays it through the interface. The report includes the patient's clinical data, risk analysis, and recommended treatment options to assist in clinical decision-making; j. The system also has historical data recording and analysis functions, which can conduct retrospective analysis of past data and provide trend forecasts for clinicians to help monitor changes in patients' health status.

3. The dynamic prediction algorithm for monitoring late-onset infections in premature infants according to claim 1, characterized in that: The Lasso regression analysis model constrains the regression coefficients through L1 regularization, so that the coefficients of unimportant variables tend to zero, thereby achieving variable selection and model optimization.

4. A system for implementing the algorithm according to claim 1 or 2, characterized in that: include: Data collection module, used to automatically collect birth weight, gestational age, 1-minute and 5-minute Apgar scores, perfusion index, categorical independent variables and other relevant clinical data of premature infants; The risk calculation module is used to input the 17 collected independent variables into the Lasso regression analysis model to calculate the regression coefficient and obtain the infection risk index of neonatal late-onset sepsis; The alarm module automatically issues a red alert when the calculated infection risk index exceeds the set threshold, reminding clinical medical staff that the newborn may be at risk of infection; A display module is used to display the calculated risk index in real time and update the display content based on clinical data and risk index; The storage module is used to store all collected clinical data, risk calculation results and historical data, and can generate reports or trend analysis for subsequent evaluation and decision-making.

5. The dynamic prediction algorithm upgrade system for monitoring late-onset infections in premature infants according to claim 4 is characterized in that: Also includes: A real-time update module is used to update the input independent variable data according to the patient's clinical status and recalculate the infection risk index; Data processing module, used to clean and normalize clinical data collected from different data sources to ensure data quality and consistency; Model optimization module, used to adjust the parameters of the Lasso regression model based on historical data and clinical feedback to improve the accuracy of prediction; The risk trend analysis module is used to display the infection risk trend of each child at different time points, assisting doctors in judging changes in the child's condition.

6. The dynamic prediction algorithm upgrade system for monitoring late-onset infections in premature infants according to claim 4 is characterized by: The alarm module further comprises: Multi-level alert function: when the infection risk index reaches different thresholds, the system can issue yellow, orange and red alerts in sequence, and the alert level is graded according to the severity of the risk; Alarm response function: after the system issues an alarm, it can guide clinical medical staff to the patient information interface and provide detailed diagnostic suggestions and treatment plans.

7. The dynamic prediction algorithm upgrade system for monitoring late-onset infections in premature infants according to claim 4 is characterized by: The formula for the L1 regularization term of the Lasso regression model is: The regularization term penalizes excessively large regression coefficients, prompting the model to select a small number of independent variables with significant influence, thereby avoiding overfitting and improving the interpretability and predictive ability of the model.

8. The dynamic prediction algorithm upgrade system for monitoring late-onset infections in premature infants according to claim 7, characterized in that: The Lasso regression model can automatically adjust the regression coefficient according to the clinical data input in real time, optimize the prediction effect and reduce the interference of unimportant variables.

Images