CN-122025184-A - Radial artery blood sampling training evaluation system and method based on big data

CN122025184ACN 122025184 ACN122025184 ACN 122025184ACN-122025184-A

Abstract

The invention belongs to the technical field of radial artery blood sampling evaluation, and particularly relates to a radial artery blood sampling training evaluation system and method based on big data. According to the invention, through multi-dimensional data acquisition and analysis, objective quantitative evaluation of radial artery blood sampling operation is realized, behavior data such as wrist joint movement, puncture angle and contact pressure of an operator can be captured, hemodynamic parameters of a patient can be synchronously acquired, dynamic characteristics in an operation process are comprehensively reflected through construction of a time-space associated data set, operation quality can be quantified through coupling analysis of behaviors and physiological responses, evaluation threshold of historical qualified samples is combined, continuity and fluency of actions are combined, evaluation is performed from the angle of an operation flow, comprehensive operation scores are generated by summarizing the operation threshold, training feedback grades are mapped, evaluation scientificity and pertinence are realized, the defect that traditional training relies on subjective experience is overcome, and objective evaluation feedback is provided.

Inventors

LIANG YAHONG
WANG JIA
WANG WENJUAN
DU YAN
ZHANG YU
LI GE
YU XUAN
WANG LEI
ZHOU YIXIAN
HUANG JING

Assignees

中国人民解放军空军军医大学

Dates

Publication Date: 20260512
Application Date: 20260210

Claims (10)

1. A radial artery blood sampling training evaluation method based on big data is characterized by comprising the following steps: Acquiring the wrist joint angular speed, the puncture needle inclination angle, the skin contact pressure distribution and the local hemodynamic parameters of a patient of an operator to form a time-space correlation data set; Extracting multidimensional features from the space-time associated data set to construct an action mode vector of an operator and a physiological response vector of a patient; Performing coupling analysis on the behavior mode vector and the physiological response vector to generate an operation quality evaluation matrix, setting an evaluation threshold value based on the historical qualified operation sample, quantifying the operation quality evaluation matrix through the evaluation threshold value, and outputting a first evaluation score; Acquiring action track data of an operator in real time, identifying action continuity and action break points of the operator based on the action track data, and determining a second evaluation score based on time distribution characteristics of the action continuity and the action break points; And carrying out weighted fusion on the first evaluation score and the second evaluation score to obtain a comprehensive operation score, and matching and outputting the training feedback level of the operator according to the comprehensive operation score.
2. The method for radial artery sampling training assessment based on big data of claim 1, wherein the step of forming the spatiotemporal association data set comprises: Acquiring the triaxial angular velocity data of the wrist joint of an operator in real time, and realizing time sequence alignment with the inclination angle data of the puncture needle through isomorphic time stamps; coordinate transformation processing is carried out on the puncture needle inclination angle data, and the puncture angle data is mapped to a human anatomy coordinate system to obtain standardized puncture angle representation; Carrying out space normalization processing on skin contact pressure distribution data, extracting the evolution and time domain characteristics of a pressure center track and a peak pressure area, and constructing a multi-mode physiological response synchronous mark sequence by combining the change time sequence of local hemodynamic parameters of a patient; and performing time alignment on the multi-mode physiological response synchronous mark sequence and the behavior operation event to form a time-space association data set.
3. The method for radial artery blood sampling training and assessment based on big data as claimed in claim 1, wherein the step of performing multi-dimensional feature extraction on the time-correlated data set to construct an operator behavior pattern vector and a patient physiological response vector comprises the steps of: Extracting frequency domain features and time domain statistics of the angular velocity of the wrist joint, and constructing a first feature sub-vector of the operation action by combining the angle change rate and acceleration features of the puncture needle; Extracting a spatial gradient characteristic and a time evolution curve of skin contact pressure distribution, fusing the displacement speed and stability indexes of the pressure center, and constructing a second characteristic sub-vector of operation contact behavior; Fusing the first characteristic sub-vector and the second characteristic sub-vector to generate a behavior pattern vector representing the dynamic characteristics of the operation behavior; the pulse waveform variation rate and the blood pressure fluctuation sequence in the hemodynamic parameters of the extractor are combined with the time alignment response of the skin pressure distribution to construct a physiological response vector of the physiological response.
4. The method for radial artery blood collection training assessment based on big data of claim 1, wherein the step of performing coupling analysis on the behavior pattern vector and the physiological response vector to generate the operation quality assessment matrix comprises the steps of: Performing time sequence alignment on the behavior pattern vector and the physiological response vector to form a unified feature map; calculating a cross correlation coefficient of the behavior mode vector and the physiological response vector in the puncture stage after the time sequence alignment, and representing the correlation strength between the operation behavior and the physiological response by using a cross correlation coefficient matrix; Based on the cross correlation coefficient matrix, identifying a synchronous occurrence period of abnormal behavior and physiological stress in the operation process, and marking the synchronous occurrence period as a key event segment; and extracting a behavior mode vector and a physiological response vector in the key event fragment, and constructing an operation quality assessment matrix for reflecting blood sampling quality.
5. The method for radial artery blood collection training assessment based on big data of claim 1, wherein the step of setting an assessment threshold based on the historical qualified operation samples, quantifying an operation quality assessment matrix by the assessment threshold, and outputting a first assessment score comprises the steps of: Extracting a compliance behavior pattern vector and a reference physiological response vector from the historical qualified operation sample, and performing cluster analysis on the compliance behavior pattern vector and the reference physiological response vector to obtain a typical operation feature distribution interval; determining upper and lower limit thresholds of each dimension characteristic based on the distribution interval, and determining a multidimensional evaluation threshold matrix; And comparing the operation quality evaluation matrix with the multidimensional evaluation threshold matrix item by item, counting the time length duty ratio and the deviation degree of each dimension characteristic exceeding the threshold interval, and outputting a first evaluation score by weighting and summing the time length duty ratio and the deviation degree of each dimension characteristic.
6. The method for radial artery blood collection training assessment based on big data of claim 1, wherein the step of identifying the action continuity and the action break point of the operator based on the action trajectory data and determining the second assessment score based on the time distribution characteristics of the action continuity and the action break point comprises the steps of: performing time domain segmentation on the action track data of an operator, and identifying action intervals corresponding to puncture preparation, needle insertion, blood collection and needle withdrawal phases; According to the acceleration variance and the speed mutation frequency of the action track in each action interval, identifying a non-fluent operation segment; Identifying a zero-speed interval in the motion track data, judging as a to-be-judged break point, and judging whether the to-be-judged break point is in an allowable transition zone between adjacent motion intervals according to whether the to-be-judged break point is in the allowable transition zone between the adjacent motion intervals; if the break point to be determined is located in the allowable transition area, determining that the normal transition is suspended, and marking the break point as the break; if the break point to be determined is located outside the allowable transition area, marking the break point as abnormal break, and recording the break point as an action break point; and counting the occurrence frequency of the action break points and the duration time proportion of the non-fluent operation fragments, and weighting to generate a second evaluation score.
7. The method for radial artery blood sampling training and assessment based on big data as set forth in claim 1, wherein the step of weighting and fusing the first assessment score and the second assessment score to obtain a comprehensive operation score, and matching and outputting a training feedback level of the operator according to the comprehensive operation score comprises the steps of: acquiring a first evaluation score and a second evaluation score, and carrying out normalization processing on the first evaluation score and the second evaluation score; respectively giving weight coefficients to the normalized first evaluation score and the normalized second evaluation score, and carrying out weighted summation by combining the normalized first evaluation score and the normalized second evaluation score to output a comprehensive operation score; mapping the comprehensive operation score with a preset segmented interval to determine a corresponding training feedback grade, wherein a plurality of segmented intervals are arranged, and each segmented interval corresponds to one training feedback grade respectively.
8. The radial artery sampling training evaluation method based on big data of claim 7, wherein the weight coefficients of the first evaluation score and the second evaluation score are adjusted according to a training phase, the weight of the second evaluation score is larger as the training phase is closer to the initial phase, and the weight of the first evaluation score is larger as the training phase is closer to the later phase.
9. A radial artery blood sampling training evaluation system based on big data, characterized in that the radial artery blood sampling training evaluation method based on big data according to any one of claims 1 to 8 is used, comprising: The data acquisition module is used for acquiring the wrist joint angular speed, the puncture needle inclination angle, the skin contact pressure distribution and the local hemodynamic parameters of the patient of an operator to form a time-space correlation data set; The feature extraction module is used for carrying out multidimensional feature extraction on the space-time associated data set and constructing a behavior pattern vector of an operator and a physiological response vector of a patient; The first evaluation module is used for carrying out coupling analysis on the behavior mode vector and the physiological response vector, generating an operation quality evaluation matrix, setting an evaluation threshold value based on the historical qualified operation samples, quantifying the operation quality evaluation matrix through the evaluation threshold value and outputting a first evaluation score; the second evaluation module is used for collecting action track data of an operator in real time, identifying action continuity and action break points of the operator based on the action track data, and determining a second evaluation score based on time distribution characteristics of the action continuity and the action break points; And the comprehensive evaluation module is used for carrying out weighted fusion on the first evaluation score and the second evaluation score to obtain a comprehensive operation score, and matching and outputting the training feedback grade of the operator according to the comprehensive operation score.
10. An electronic device, characterized in that the electronic device comprises: At least one processor; and a memory communicatively coupled to the at least one processor; Wherein the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the big data based radial arterial blood collection training assessment method of any one of claims 1 to 8.

Description

Radial artery blood sampling training evaluation system and method based on big data Technical Field The invention belongs to the technical field of radial artery blood sampling evaluation, and particularly relates to a radial artery blood sampling training evaluation system and method based on big data. Background Radial artery blood sampling is used as a common clinical invasive operation, the technological mastering degree of the radial artery blood sampling directly influences the blood sampling success rate and the patient experience, so that necessary puncture training is obviously necessary before medical staff formally carrying out clinical operation, thereby not only improving the operation standardization, but also effectively reducing the complication risk caused by unskilled technology, and simultaneously providing corresponding safety guarantee and comfortable experience for the patient, so as to reduce the pain and vascular injury related to puncture. In the prior art, radial artery blood sampling training is mostly dependent on a traditional teaching mode, namely on-site observation and subjective experience evaluation by a mentor, so that the mode lacks objective and quantitative evaluation standards, is difficult to intuitively reflect the real skill level of an operator, is easily interfered by subjective factors, and further can lead to inconsistent evaluation results and influence the scientificity of training effects. Disclosure of Invention The invention aims to provide a radial artery blood sampling training evaluation system and method based on big data, which can realize quantitative evaluation of puncture actions of operators through multi-dimensional data acquisition and analysis and make up for the defects of strong subjectivity and feedback lag in the traditional training. The technical scheme adopted by the invention is as follows: A radial artery blood sampling training evaluation method based on big data comprises the following steps: Acquiring the wrist joint angular speed, the puncture needle inclination angle, the skin contact pressure distribution and the local hemodynamic parameters of a patient of an operator to form a time-space correlation data set; Extracting multidimensional features from the space-time associated data set to construct an action mode vector of an operator and a physiological response vector of a patient; Performing coupling analysis on the behavior mode vector and the physiological response vector to generate an operation quality evaluation matrix, setting an evaluation threshold value based on the historical qualified operation sample, quantifying the operation quality evaluation matrix through the evaluation threshold value, and outputting a first evaluation score; Acquiring action track data of an operator in real time, identifying action continuity and action break points of the operator based on the action track data, and determining a second evaluation score based on time distribution characteristics of the action continuity and the action break points; And carrying out weighted fusion on the first evaluation score and the second evaluation score to obtain a comprehensive operation score, and matching and outputting the training feedback level of the operator according to the comprehensive operation score. In a preferred embodiment, the step of forming a spatio-temporal associated dataset includes: Acquiring the triaxial angular velocity data of the wrist joint of an operator in real time, and realizing time sequence alignment with the inclination angle data of the puncture needle through isomorphic time stamps; coordinate transformation processing is carried out on the puncture needle inclination angle data, and the puncture angle data is mapped to a human anatomy coordinate system to obtain standardized puncture angle representation; Carrying out space normalization processing on skin contact pressure distribution data, extracting the evolution and time domain characteristics of a pressure center track and a peak pressure area, and constructing a multi-mode physiological response synchronous mark sequence by combining the change time sequence of local hemodynamic parameters of a patient; and performing time alignment on the multi-mode physiological response synchronous mark sequence and the behavior operation event to form a time-space association data set. In a preferred embodiment, the step of performing multidimensional feature extraction on the time-space correlation data set to construct a behavior pattern vector of an operator and a physiological response vector of a patient includes: Extracting frequency domain features and time domain statistics of the angular velocity of the wrist joint, and constructing a first feature sub-vector of the operation action by combining the angle change rate and acceleration features of the puncture needle; Extracting a spatial gradient characteristic and a time evolution curve of skin contact pressure