WO2017097040A1

WO2017097040A1 - Method and system for evaluating medical transfusion speed

Info

Publication number: WO2017097040A1
Application number: PCT/CN2016/102671
Authority: WO
Inventors: 刘宇航; 聂泽东; 李景振; 王磊
Original assignee: 深圳先进技术研究院
Priority date: 2015-12-10
Filing date: 2016-10-20
Publication date: 2017-06-15
Also published as: CN105550509B; CN105550509A

Abstract

Provided are a method and system for evaluating a medical transfusion speed, comprising: determining an initial weight value according to an impact factor; determining a weight value of the impact factor according to the initial weight value; and then determining an evaluation result according to the weight value, thereby implementing the method for evaluating the safety and drug efficacy of a transfusion speed according to a population, disease and drug, and facilitating people's intuitive understanding of the safety and drug efficacy of a transfusion speed.

Description

Medical transport droplet velocity evaluation method and system

Technical field

The present invention relates to the field of medical technology, and in particular, to a medical infusion rate evaluation method and system.

Background technique

With the development of the economy, people pay more and more attention to health, and the factors that pose risks to their health are more concerned. However, because the infusion flow rate is improper, the body's body is uncomfortable, and it is life-threatening. The determination of the infusion flow rate is basically determined by the medical staff based on clinical experience and personal opinion. The lack of a reliable assessment standard is to test whether the infusion flow rate is safe and effective. There are many factors affecting the safety of infusion flow rate, such as the type of drug, the type of disease and the age of the patient, the infusion flow rate, how to determine the degree of influence of different factors and the weight is the key to the safety evaluation of the infusion flow rate. The weight distribution of various factors uses BP neural network feedback algorithm, which is simple and accurate, so that the weight of the influencing factors can be accurately determined. Due to the diversity of impact factors, it is necessary to adopt comprehensive evaluation to measure whether the weight distribution of each factor is reasonable, so as to determine whether the infusion flow rate under different factors is safe and effective, and provide protection for people's health.

The method of determining the weight of influencing factors by neural network is one of the most widely used and well-established methods of weighting. Among many neural network algorithms, BP neural network algorithm is the most widely used, and the application effect is the best. . In other applications of BP neural network learning algorithm, the initial weights are randomly set to a small number, and the initial weight of the technology is given by medical experts, which greatly shortens the network training time. In addition, because there are many factors affecting the infusion flow rate, when evaluating the infusion flow rate, all factors must be considered at the same time, and comprehensive evaluation is needed. In most cases, it is difficult to judge with a simple numerical value, so it is better to use fuzzy comprehensive evaluation of the infusion flow rate. The actual effect.

At present, the medical staff's control of the infusion flow rate is determined by experience. The lack of accurate standard measurement determines whether the infusion flow rate is safe and effective for the human body. If the flow rate is improperly controlled, the patient's life is endangered. At the same time, the effects of drugs, diseases, age and other factors on the flow rate are not comprehensively considered. The infusion flow rate is often determined only according to a single factor with the greatest influence. However, such flow rate control cannot guarantee that the safety of the drug to the human body is optimal, and it cannot be guaranteed. The efficacy of the drug being delivered plays an optimal role.

Summary of the invention

Based on this, the present invention provides a medical infusion rate evaluation method to achieve the maximum effect of infusion drugs on human safety and efficacy.

The invention adopts the following technical solutions:

A method for evaluating a medical drop velocity includes the following steps:

Determine the impact factor;

Determining an initial weight value according to the impact factor;

Determining a weight value of the impact factor according to the weight initial value; and

The evaluation result is determined based on the weight value.

In some embodiments, the influencing factors are divided into drugs, diseases, ages, and drip speeds, and subdivided on the basis, the drugs are divided into colloids, dehydrating agents, antibiotics, vasoactive drugs, and others, and the diseases are divided into hearts. , within the respiratory, intra-digestive, intratumoral, other internal, extrathoracic, extracerebral, urinary, the population is divided into less than 1 year old, 1 to 3 years old, 4 to 12 years old, 13 to 18 years old, 19 to 60 years old, greater than At the age of 60, the drip rate is divided into less than 40 drops/min, 40 to 60 drops/min, 61 to 80 drops/min, and 81 to 120 drops/min.

In some embodiments, wherein determining the weight value of the impact factor according to the weight initial value is specifically: determining an impact factor weight value by a neural network algorithm, where the neural network is a three-layer neural network Network, including input layer, middle layer, and output layer.

In some embodiments, the number of neurons of the hidden layer is greater than half of a sum of the number of neurons of the input layer neurons and the output layer, less than the number of neurons of the input layer neurons and the output layer And.

In some embodiments, the impact factor weight value is determined by a neural network algorithm, including the following steps:

Constructing the three-layer neural network parameter and the droplet velocity evaluation system;

Select the appropriate amount of samples to use the particle swarm optimization algorithm for BP network training;

The influence factor weight value is determined according to the training result.

In some embodiments, determining the impact factor weight value based on the training result includes:

Constructing a first formula, a second formula, and a third formula, respectively, the first formula being:

The second formula is:

The third formula is:

Where i is the neural network input unit, i=1,...m;j is the neural network output unit, j=1,...n; k is the implicit unit of the neural network, k=1,... p; ki is the weight coefficient between the input layer neuron i and the hidden layer neuron k;

The influence factor weight value S is obtained according to the first formula, the second formula, and the third formula.

In some embodiments, determining the evaluation result based on the weight value includes the following steps:

1. Set the factor set:

U=(u ₁ , u ₂ , u ₃ , u ₄ , u ₅ , u ₆ , u ₇ , u ₈ , u ₉ , u ₁₀ , u ₁₁ , u ₁₂ , u ₁₃ , u ₁₄ , u ₁₅ , u ₁₆ , u ₁₇ , u ₁₈ , u ₁₉ , u ₂₀ , u ₂₁ , u ₂₂ , u ₂₃ ) wherein u ₁ is a colloid, u ₂ is a dehydrating agent, u ₃ is an antibiotic, u ₄ is a vasoactive drug, and u ₅ is other Drugs, u ₆ for the heart, u ₇ for the respiratory, u ₈ for the digestive, u _{9 for the} tumor, u ₁₀ for the other, u ₁₁ for the chest, u ₁₂ for the brain, u _{13 for the} urinary, u ₁₄ For <1 year old, u ₁₅ is 1-3 years old, u ₁₆ is 4-12 years old, u ₁₇ is 13-18 years old, u ₁₈ is 19-60 years old, u ₁₉ is >60 years old, u ₂₀ is <40 drops /min, u ₂₁ is 40-60 drops / min, u ₂₂ is 61-80 drops / min, u ₂₃ is 81-120 drops / min;

2. Set the evaluation set:

V={v ₁ , v ₂ , v ₃ , v ₄ , v ₅ , v ₆ , v ₇ , v ₈ , v ₉ }

Among them, v ₁ is excellent in safety and excellent in efficacy, v ₂ is excellent in safety and good in efficacy, v ₃ is excellent in safety and good in efficacy, v ₄ is good in safety and excellent in efficacy, and v ₅ is safe. Good and effective, v ₆ is safe and effective, v ₇ is safe and excellent, v ₈ is safe and good, v ₉ is safe and general;

3. Establish a judging matrix, that is, establish a fuzzy mapping from U to F(V):

f: U → F (V),

The fuzzy relation R can be induced by f, and the single factor evaluation matrix is obtained.

4. Determine the allocation of weights:

5. Comprehensive evaluation:

After R is obtained from A, the comprehensive evaluation is B=AoR, and B={b ₁ , b ₂ ,..., b _m }, where B is a fuzzy subset on V, where

If judgement result

Then normalize b _j , and finally according to the principle of maximum membership, the item corresponding to the maximum b _j is the result of the judgment.

In addition, the present invention also provides an evaluation system for medical droplet velocity, comprising:

An impact factor determination module for determining an impact factor;

a weight initial value determining module, configured to determine a weight initial value according to the impact factor;

a weight value determining module, configured to determine a weight value of the impact factor according to the weight initial value; and

And an evaluation module, configured to determine an evaluation result according to the weight value.

The method and system for evaluating the medical drop velocity according to the present invention, determining an initial value of the weight according to the influence factor, determining a weight value of the influence factor according to the initial value of the weight, and determining an evaluation result according to the weight value, thereby realizing According to the population, diseases, drugs to measure the speed and efficacy of the drop speed, it is convenient for people to have a more intuitive understanding of the safety and efficacy of the drop speed.

In addition, the above technical solution uses the neural network algorithm to redistribute the influence factor weights, and then obtains reasonable and scientific weights. At the same time, the fuzzy comprehensive evaluation method is used to evaluate the facts according to the influence factor weights, and then a scientific evaluation result is obtained. In order to achieve the maximum effect of infusion drugs on human safety and efficacy, guide medical staff to control the infusion flow rate more accurately, so that the risk of affecting people's health is reduced, and people's health is guaranteed.

DRAWINGS

1 is a flow chart showing the steps of a method for evaluating a medical drop rate according to a preferred embodiment of the present invention.

2 is a schematic diagram of an impact factor according to a preferred embodiment of the present invention.

FIG. 3 is a flowchart of an IPSO optimized network training algorithm according to a preferred embodiment of the present invention.

4 is a medical infusion rate evaluation system provided by the present application.

detailed description

In order to facilitate the understanding of the present invention, the present invention will be described more fully hereinafter with reference to the accompanying drawings. Preferred embodiments of the invention are given in the drawings. The above are only the preferred embodiments of the present invention, and are not intended to limit the scope of the invention, and the equivalent structure or equivalent process transformations made by the description of the present invention and the drawings are directly or indirectly applied to other related technical fields. The same is included in the scope of patent protection of the present invention.

All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. The terminology used in the description of the present invention is for the purpose of describing particular embodiments and is not intended to limit the invention. The term "and/or" used herein includes any and all combinations of one or more of the associated listed items.

Referring to FIG. 1 , a preferred embodiment of the present invention provides a medical drop rate evaluation method 100, including:

Step S110: determining an impact factor;

According to clinical experience, the impact factors are divided into drugs, diseases, ages and drip rates, which are subdivided on this basis. The drug is divided into colloid, dehydrating agent, antibiotic, vasoactive drug and others. The disease is divided into intracardiac, intragastric, intragastric, intratumoral, other internal, extrathoracic, extracerebral, and urinary. The population is divided into less than 1 year old. 1 to 3 years old, 4 to 12 years old, 13 to 18 years old, 19 to 60 years old, more than 60 years old, the drip rate is divided into less than 40 drops / min, 40 to 60 drops / min, 61 to 80 drops / min, 81 to 120 drops / min. The impact factor is shown in Figure 2.

Step S120: determining an initial weight value according to the impact factor;

In practice, determining the initial weight value according to the influence factor is given a preliminary weight by a doctor with rich clinical experience, and then scientifically demonstrating the result of the weight distribution by the experts in the medical field, and finally determining the distribution of the initial weight value.

Step S130: determining a weight value of the impact factor according to the weight initial value;

Preferably, the impact factor weight value is determined by a neural network algorithm that is a three-layer neural network including an input layer, an intermediate layer, and an output layer.

Among them, the input layer is 21 index layers, the middle layer is taken as one in the current calculation, and the output layer is one, which is the weight value of each influencing factor. A single output BP neural network topology table that measures the safety and efficacy of droplet velocity is shown in Table 1.

Table 1 Single Output BP Neural Network Topology Table for Measuring Droplet Velocity Safety and Pharmacodynamics

Specifically, the present application determines the impact factor weight value by using a neural network algorithm, including the following steps:

Step S131: constructing the three-layer neural network parameter and the drop velocity evaluation system;

Specifically, it is evaluated from four aspects: population, disease, drug and drip rate. The indicators are as shown in Table 2. The factor is shown.

Table 2 Specific parameters of the safety and pharmacodynamics of the droplet velocity

It can be understood that the number of hidden layer neural units in the above table can be set by itself. Generally speaking, if the problem to be solved is more complicated, the number of hidden layer units should be set more, or the same problem, the more hidden layers are. The easier it is to converge, but if you set too many hidden layer units to increase the amount of calculation, there is currently no effective method for setting the number of hidden layer units. Generally, it needs to be determined according to the size of the network.

The present application determines the number of cells of the hidden layer according to the following rules: the number of neurons in the hidden layer is greater than half of the sum of the number of neurons in the input layer and the output layer, and is smaller than the sum of the number of neurons in the input layer and the output layer. .

Step S132: selecting an appropriate amount of samples by using a particle swarm optimization algorithm for training of the BP network;

It can be understood that after the network topology is established, an appropriate amount of samples need to be selected to train and learn the network. This application initializes the sample data by the initial value of the weight given by the medical field expert. From the principle of BP neural network algorithm, the larger the number of samples, the better, but the appropriate sample size should be determined according to the size of the network. If the size is too large or too small, the calculation will be inaccurate. After completing this step, the training results of the neural network will be obtained. .

Please refer to FIG. 3 , which is a flowchart of an IPSO optimized network training algorithm according to an embodiment of the present invention. It can be understood that the particle swarm optimization algorithm (IPOS ) is used for training of a BP network, that is, the position of each particle in a particle group is represented in a BP network. The set of weights of the current iteration, the dimension of each particle is connected by the network The number of weights and the number of thresholds are determined. The neural network output error of a given training sample set is used as an adaptive function of the neural network training problem. The fitness value represents the error of the neural network. The smaller the error, the better the performance of the particle in the search. Moving the search within the weight space minimizes the error in the output layer of the network. Changing the speed of the particle updates the weight of the network to reduce the mean square error (MSE).

Step S133: Determine the influence factor weight value according to the training result.

It can be understood that the purpose of establishing a neural network learning algorithm is to determine the weight value of the impact factor, and the result of the neural network training is only the relationship between the neurons, if you want to get the true relationship between the input factor and the output factor, That is, the weight of the input factors affects the output factors, and the weights between the neurons need to be analyzed and processed. Therefore, the following indicators are used to describe the relationship between input factors and output factors.

1 correlation significance coefficient

2 related index

3 absolute influence coefficient

In the above formula: i is the neural network input unit, i=1,...m; j is the neural network output unit, j=1,...n; k is the implicit unit of the neural network, k=1,. ..p;wki is the weight coefficient between the input layer neuron i and the hidden layer neuron k.

The absolute influence coefficient S among the above three correlation coefficients is the required weight value. Applying formula (1) to (3) The weight values of the respective influence factors can be calculated.

Step S140: Determine an evaluation result according to the weight value.

The fuzzy comprehensive evaluation method is used to judge a certain fact according to the weight value of the influence factor determined above. The specific process is as follows:

1: Set factor set

U=(u ₁ , u ₂ , u ₃ , u ₄ , u ₅ , u ₆ , u ₇ , u ₈ , u ₉ , u ₁₀ , u ₁₁ , u ₁₂ , u ₁₃ , u ₁₄ , u ₁₅ , u ₁₆ , u ₁₇ , u ₁₈ , u ₁₉ , u ₂₀ , u ₂₁ , u ₂₂ , u ₂₃ ) wherein u ₁ is a colloid, u ₂ is a dehydrating agent, u ₃ is an antibiotic, u ₄ is a vasoactive drug, and u ₅ is other Drugs, u ₆ for the heart, u ₇ for the respiratory, u ₈ for the digestive, u _{9 for the} tumor, u ₁₀ for the other, u ₁₁ for the chest, u ₁₂ for the brain, u _{13 for the} urinary, u ₁₄ For <1 year old, u ₁₅ is 1-3 years old, u ₁₆ is 4-12 years old, u ₁₇ is 13-18 years old, u ₁₈ is 19-60 years old, u ₁₉ is >60 years old, u ₂₀ is <40 drops /min, u ₂₁ is 40-60 drops/min, u ₂₂ is 61-80 drops/min, and u ₂₃ is 81-120 drops/min.

2: Set the evaluation set

V={v ₁ , v ₂ , v ₃ , v ₄ , v ₅ , v ₆ , v ₇ , v ₈ , v ₉ }

Among them, v ₁ is excellent in safety and excellent in efficacy, v ₂ is excellent in safety and good in efficacy, v ₃ is excellent in safety and good in efficacy, v ₄ is good in safety and excellent in efficacy, and v ₅ is safe. Good and effective, v ₆ is safe and effective, v ₇ is safe and excellent, v ₈ is safe and good, v ₉ is safe and general.

3: Establish a judgment matrix. That is to establish a fuzzy mapping from U to F (V)

f: U → F (V),

4. Determine the weight. Since each factor in U has different emphasis, each factor needs to be given different weights, which can be expressed as a fuzzy subset A on U={a ₁ , a ₂ ,..., a _n }, and Regulation

The weights in this paper are given preliminary weights by doctors with rich clinical experience, and then scientifically demonstrated by the experts in the medical field to determine the weight distribution.

5 comprehensive evaluation. After R is obtained from A, the comprehensive evaluation is B = AoR, and B = {b ₁ , b ₂ , ..., b _m } which is a fuzzy subset on V. among them

If judgement result

It should be normalized. Finally, according to the principle of maximum membership, the item corresponding to the maximum b _j is the result of the judgment.

Referring to FIG. 4, the present application further provides a medical infusion rate evaluation system, including: an impact factor determining module 110, configured to determine an impact factor; a weight initial value determining module 120, configured to determine a weight initial according to the influencing factor a weight value determining module 130, configured to determine a weight value of the impact factor according to the weight initial value; and an evaluation module 140, configured to determine an evaluation result according to the weight value. See the description above for details.

In addition, the above technical solution uses a neural network algorithm to redistribute the influence factor weights into The rational and scientific weights are obtained. At the same time, the fuzzy comprehensive evaluation method is used to evaluate the facts according to the weight of the impact factors, and then a scientific evaluation result is obtained to realize the maximum effect of the infusion drugs on the safety and efficacy of the human body. Instruct medical staff to control the infusion flow rate more accurately, so that the risk of affecting people's health is reduced, and people's health is guaranteed.

The above-mentioned embodiments are merely illustrative of several embodiments of the present invention, and the description thereof is more specific and detailed, but is not to be construed as limiting the scope of the invention. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A method for evaluating a medical droplet velocity, comprising the steps of:

Determine the impact factor;

Determining an initial weight value according to the impact factor;

Determining a weight value of the impact factor according to the weight initial value; and

The evaluation result is determined based on the weight value.
The medical drip rate evaluation method according to claim 1, wherein the influence factor is classified into a drug, a disease, an age, and a drip rate, wherein the drug is classified into a colloid, a dehydrating agent, an antibiotic, a vasoactive drug, and the like. The disease is divided into intracardiac, intragastric, intragastric, intratumoral, other internal, extrathoracic, extracerebral, and urinary. The population is divided into less than 1 year old, 1 to 3 years old, 4 to 12 years old, 13 to 18 years old, 19 to 60 years old, more than 60 years old, the drip rate is divided into less than 40 drops / min, 40 to 60 drops / min, 61 to 80 drops / min, 81 to 120 drops / min.
The method according to claim 1, wherein determining the weight value of the influence factor according to the weight initial value is specifically: determining a influence factor weight value by a neural network algorithm, The neural network is a three-layer neural network including an input layer, an intermediate layer, and an output layer.
The method according to claim 3, wherein the number of neurons of the hidden layer is greater than half of a sum of the number of neurons of the input layer and the output layer, less than The sum of the number of input layer neurons and the output layer neurons.
The medical droplet velocity evaluation method according to claim 3, wherein the influence factor weight value is determined by a neural network algorithm, comprising the following steps:

Constructing the three-layer neural network parameter and the droplet velocity evaluation system;

Select the appropriate amount of samples to use the particle swarm optimization algorithm for BP network training;

The influence factor weight value is determined according to the training result.
The medical drop velocity evaluation method according to claim 5, wherein determining the influence factor weight value according to the training result comprises the following steps:

Constructing a first formula, a second formula, and a third formula, respectively, the first formula being:

The second formula is:

The third formula is:

Where i is the neural network input unit, i=1,...m;j is the neural network output unit, j=1,...n; k is the implicit unit of the neural network, k=1,... p; ki is the weight coefficient between the input layer neuron i and the hidden layer neuron k;

The influence factor weight value S is obtained according to the first formula, the second formula, and the third formula.
The medical drip rate evaluation method according to claim 1, wherein determining the evaluation result based on the weight value comprises the following steps:

1. Set the factor set:

U=(u 1 , u 2 , u 3 , u 4 , u 5 , u 6 , u 7 , u 8 , u 9 , u 10 , u 11 , u 12 , u 13 , u 14 , u 15 , u 16 , u 17 , u 18 , u 19 , u 20 , u 21 , u 22 , u 23 ) wherein u 1 is a colloid, u 2 is a dehydrating agent, u 3 is an antibiotic, u 4 is a vasoactive drug, and u 5 is other Drugs, u 6 for the heart, u 7 for the respiratory, u 8 for the digestive, u 9 for the tumor, u 10 for the other, u 11 for the chest, u 12 for the brain, u 13 for the urinary, u 14 For <1 year old, u 15 is 1-3 years old, u 16 is 4-12 years old, u 17 is 13-18 years old, u 18 is 19-60 years old, u 19 is >60 years old, u 20 is <40 drops /min, u 21 is 40-60 drops / min, u 22 is 61-80 drops / min, u 23 is 81-120 drops / min;

2. Set the evaluation set:

V={v 1 , v 2 , v 3 , v 4 , v 5 , v 6 , v 7 , v 8 , v 9 }

Among them, v 1 is excellent in safety and excellent in efficacy, v 2 is excellent in safety and good in efficacy, v 3 is excellent in safety and good in efficacy, v 4 is good in safety and excellent in efficacy, and v 5 is safe. Good and effective, v 6 is safe and effective, v 7 is safe and excellent, v 8 is safe and good, v 9 is safe and general;

3. Establish a judging matrix, that is, establish a fuzzy mapping from U to F(V):

f: U → F (V),

The fuzzy relation R can be induced by f, and the single factor evaluation matrix is obtained.

4. Determine the allocation of weights:

5. Comprehensive evaluation:

After R is obtained from A, the comprehensive evaluation is B=AoR, and B={b 1 , b 2 ,..., b m }, where B is a fuzzy subset on V, where
If judgement result
Then normalize b j , and finally according to the principle of maximum membership, the item corresponding to the maximum b j is the result of the judgment.
An evaluation system for medical drop velocity, characterized in that it comprises:

An impact factor determination module for determining an impact factor;

a weight initial value determining module, configured to determine a weight initial value according to the impact factor;

a weight value determining module, configured to determine a weight value of the impact factor according to the weight initial value; And an evaluation module, configured to determine an evaluation result according to the weight value.