CN-121983235-A - Digital twin model-based neonatal care plan simulation method and system

CN121983235ACN 121983235 ACN121983235 ACN 121983235ACN-121983235-A

Abstract

The invention provides a neonate nursing scheme simulation method and system based on a digital twin model, wherein the method comprises the steps of carrying out nursing scheme gene coding on extracted neonate historical nursing data to obtain a nursing scheme gene library, generating candidate nursing schemes based on the nursing scheme gene library, utilizing an evolution algorithm to input the candidate nursing schemes into the digital twin model, introducing random noise into the digital twin model to simulate physiological fluctuation of a neonate to generate state distribution data, constructing opposite comparison scenes for each candidate nursing scheme, keeping other conditions unchanged, only adjusting preset nursing parameters, obtaining a state difference value through calculation of neonate state track differences under different nursing parameters, calculating comprehensive evaluation indexes based on the state distribution data and the state difference value, and further screening out an optimal nursing scheme to identify a nursing scheme which is stable in performance and high in accuracy in an uncertain environment so as to meet clinical requirements of neonate nursing on refined and personalized decisions.

Inventors

XU YUEYUE
HU YUEMING
YU CHUNXIA
LIAO QUANXING

Assignees

泓禾正(广州)医疗设备有限公司

Dates

Publication Date: 20260505
Application Date: 20251231

Claims (10)

1. A method for simulating a neonatal care regimen based on a digital twin model, comprising: Extracting neonatal historical care data from a neonatal medical information system; Carrying out care scheme gene coding on the neonatal history care data to obtain a care scheme gene library; generating candidate nursing schemes by using an evolution algorithm based on the nursing scheme gene library; inputting the candidate nursing scheme into a digital twin model, introducing random noise into the digital twin model to simulate physiological fluctuation of the neonate, and generating state distribution data reflecting the state track of the neonate; Constructing opposite comparison scenes for each candidate nursing scheme, keeping other conditions unchanged, only adjusting preset nursing parameters, and obtaining a state difference value through calculating the neonatal state track difference under different comparison nursing parameters; And calculating a comprehensive evaluation index reflecting the nursing effect of the candidate nursing scheme based on the state distribution data and the state difference value, and screening an optimal nursing scheme from the candidate nursing schemes based on the comprehensive evaluation index.
2. The method of claim 1, wherein the neonatal historic care data comprises historic care regimen and care parameters thereof, neonatal physiological data, biochemical data, therapeutic data, and developmental data.
3. The method of claim 1, wherein said encoding of the care plan genes of the neonatal historic care data results in a care plan gene library comprising: Extracting care parameters from the neonatal historic care data; Dividing the nursing parameters into numerical parameters and category parameters; Dividing the numerical parameters into intervals, and mapping the numerical value of each numerical parameter to a corresponding numerical interval; Establishing a numerical value set for the category type parameters, and mapping the numerical value of each category type parameter to a corresponding element in the set; Converting the mapped numerical value interval and the aggregation element into binary codes; According to the time sequence and the execution sequence of the nursing parameters, the converted binary codes are spliced end to obtain a nursing scheme gene sequence; and storing the care scheme gene sequence into a relational database, and constructing to obtain the care scheme gene library.
4. The method of claim 1, wherein the encoding of the care plan genes for the neonatal historic care data further comprises, prior to obtaining a care plan gene library: Preprocessing the neonatal history care data, wherein the preprocessing comprises data cleaning and data format unification.
5. The method of claim 1, wherein generating candidate care regimens using an evolutionary algorithm based on the care regimen gene library comprises: Selecting a care plan gene sequence from the care plan gene library, and constructing an initial population, wherein the initial population comprises at least three care plan gene sequences; performing reverse mapping operation on each care plan gene sequence in the initial population to obtain a corresponding care parameter combination; Evaluating the nursing effect under the nursing parameter combination to obtain an evaluation value corresponding to each nursing parameter combination; Sorting the nursing scheme gene sequences in the initial group according to each evaluation value, screening the nursing scheme gene sequences ranked at the front N positions to obtain a first nursing scheme gene sequence, wherein N is a positive integer; Pairing the first care scheme gene sequences, and exchanging coding fragments of each pair of care scheme gene sequences at the middle position of the gene sequences to generate a second care scheme gene sequence; Selecting a coding position at the rear part of the gene sequence from the second care scheme gene sequence, and replacing the current value of the coding position with an opposite value to generate a third care scheme gene sequence; Combining a third care scheme gene sequence with the first care scheme gene sequence to form a new generation group, replacing the initial group with the new generation group, and returning to execute the step of executing reverse mapping operation on each care scheme gene sequence in the initial group, and repeating the steps until the evaluation value of the care scheme gene sequence with the highest evaluation value in the group in two adjacent operations is kept consistent, so as to obtain a target care scheme gene sequence; and performing reverse mapping operation on the target care plan gene sequence to obtain candidate care plans.
6. The method of claim 5, wherein performing a reverse mapping operation on each care regimen gene sequence in the initial population results in a corresponding care parameter combination, comprising: and reversely converting each care scheme gene sequence in the initial population according to the mapping rule during coding to obtain a corresponding care parameter combination.
7. The method of claim 1, wherein calculating a composite assessment indicator reflecting the candidate care regimen care effect based on the status distribution data and status difference values comprises: constructing a time series set of neonatal status parameters based on the status distribution data; calculating the difference between the maximum value and the minimum value of the state parameter values generated in all simulation operations for each time point in the time sequence set to obtain the fluctuation amplitude of the neonatal state parameter at each time point; counting the number of target time points with the fluctuation amplitude larger than a preset safety threshold in the time sequence set, calculating the proportion of the number of the target time points to the number of the total time points, and generating a risk frequency index; and carrying out weighted summation on the risk frequency index and the state difference value to obtain a comprehensive evaluation index.
8. The method of claim 1, wherein the status distribution data comprises neonatal weight gain, height gain, head circumference gain, blood oxygen saturation, blood glucose stability and bilirubin stability.
9. The method of claim 1, wherein the predetermined care parameters include a nutrition supply parameter, a temperature and humidity of the incubator, a ventilator ventilation parameter, a phototherapy parameter, a medication injection parameter, a turn-over frequency, and a number of beats.
10. A digital twinning model-based neonatal care regimen simulation system, comprising: An extraction module for extracting neonatal historic care data from a neonatal medical information system; the coding module is used for carrying out nursing scheme gene coding on the neonatal history nursing data to obtain a nursing scheme gene library; The generation module is used for generating candidate nursing schemes by using an evolution algorithm based on the nursing scheme gene library; The input module is used for inputting the candidate nursing scheme into a digital twin model, introducing random noise into the digital twin model to simulate physiological fluctuation of the neonate, and generating state distribution data reflecting the state track of the neonate; The adjusting module is used for constructing opposite comparison scenes for each candidate nursing scheme, keeping other conditions unchanged, only adjusting preset nursing parameters, and obtaining a state difference value through calculating the neonatal state track difference under different comparison nursing parameters; and the screening module is used for calculating a comprehensive evaluation index reflecting the nursing effect of the candidate nursing scheme based on the state distribution data and the state difference value, and screening an optimal nursing scheme from the candidate nursing schemes based on the comprehensive evaluation index.

Description

Digital twin model-based neonatal care plan simulation method and system Technical Field The invention relates to the technical field of neonatal care, in particular to a neonatal care scheme simulation method and system based on a digital twin model. Background Neonatal monitoring is a special field with higher technical requirements and tighter risk management in clinical practice. Neonates, especially premature infants and critically ill infants, have physiological systems such as cardiopulmonary function, nervous system, metabolic level and the like which are not yet developed and mature, and respond to external nursing operations with obvious individual variability and state instability. Therefore, in the actual nursing process, the medical staff needs to dynamically adjust various nursing parameters such as ventilation support parameters of the breathing machine, oxygen therapy concentration, intravenous nutrition infusion rate, posture management strategy, environmental temperature and humidity control parameters and the like so as to maintain the physiological state of the neonate stable and promote organ development. However, in the current neonatal monitoring, the establishment of a nursing scheme mainly depends on standardized suggestions of clinical guidelines and the accumulation of personal experience of medical staff, and the method has the fundamental defects that on one hand, massive neonatal nursing data are stored in a medical information system in a fragmented and uncorrelated form for a long time, and cannot be converted into a structured knowledge base capable of supporting the generation or optimization of an automatic scheme, so that data resources are idle and wasted, and on the other hand, the conventional nursing scheme assessment method generally adopts a static comparison method and cannot simulate inherent random fluctuation of a neonatal physiological system, so that the nursing scheme is easily influenced by environmental noise and has low accuracy, and the clinical requirements of the neonatal nursing on fine and personalized decisions are difficult to meet. Disclosure of Invention The invention provides a digital twin model-based neonatal care scheme simulation method and system, which are used for meeting clinical requirements of neonatal care on refined and personalized decisions. In order to solve the problems, the invention adopts the following technical scheme: In a first aspect, the present invention provides a method for simulating a neonatal care regimen based on a digital twin model, comprising: Extracting neonatal historical care data from a neonatal medical information system; Carrying out care scheme gene coding on the neonatal history care data to obtain a care scheme gene library; generating candidate nursing schemes by using an evolution algorithm based on the nursing scheme gene library; inputting the candidate nursing scheme into a digital twin model, introducing random noise into the digital twin model to simulate physiological fluctuation of the neonate, and generating state distribution data reflecting the state track of the neonate; Constructing opposite comparison scenes for each candidate nursing scheme, keeping other conditions unchanged, only adjusting preset nursing parameters, and obtaining a state difference value through calculating the neonatal state track difference under different comparison nursing parameters; And calculating a comprehensive evaluation index reflecting the nursing effect of the candidate nursing scheme based on the state distribution data and the state difference value, and screening an optimal nursing scheme from the candidate nursing schemes based on the comprehensive evaluation index. Preferably, the neonatal historical care data includes a historical care regimen and care parameters thereof, neonatal physiological data, biochemical data, therapeutic data, and developmental data. Preferably, the coding of the care plan gene of the neonatal history care data to obtain a care plan gene library comprises: Extracting care parameters from the neonatal historic care data; Dividing the nursing parameters into numerical parameters and category parameters; Dividing the numerical parameters into intervals, and mapping the numerical value of each numerical parameter to a corresponding numerical interval; Establishing a numerical value set for the category type parameters, and mapping the numerical value of each category type parameter to a corresponding element in the set; Converting the mapped numerical value interval and the aggregation element into binary codes; According to the time sequence and the execution sequence of the nursing parameters, the converted binary codes are spliced end to obtain a nursing scheme gene sequence; and storing the care scheme gene sequence into a relational database, and constructing to obtain the care scheme gene library. Further, before the care plan gene encoding is performed on the neonatal hist