CN102831326A - Mean amplitude of glucose excursions (MAGE) calculation method - Google Patents

Abstract

The invention relates to a MAGE (mean amplitude of glucose excursions) calculation method. The method comprises reading a dynamic glucose monitoring data obtained from a dynamic glucose monitor instrument, and calculating the glucose standard deviation of the obtained data; finding all extreme points of the obtained data to obtain an extreme point sequence set; finding an effective extreme point according to the principle that the absolute value of the glucose difference of two adjacent extreme points in the set is not smaller than the glucose standard deviation and that the two adjacent extreme points are respectively the maximum value and the minimum value; and calculating the MAGE according to a MAGE calculation formula. The method provided by the invention can rapidly and accurately analyze the dynamic glucose monitoring data to obtain the effective parameters, and can greatly shorten the calculation time while ensuring the calculation accuracy of MAGE, thereby improving the efficiency of clinical work and scientific research.

Description

A kind of computing method of MAGE

Technical field

The present invention relates to the measurement of body inner blood characteristic, be specifically related to the disposal route of MAGE parameter in the human body.

Background technology

(Mean Amplitude of Glycemic Excursions MAGE) is a kind of referential data of the reaction blood glucose fluctuation of setting up based on dynamic glucose monitor data to MAGE.In the present practical clinical; The medical personnel lets the disposable monitoring instrument of wearing of patient obtain above 72 hours above 864 dynamic glucose monitor datas; And MAGE numerical value often can not directly directly be analyzed acquisition from dynamic blood glucose monitoring system; Still must screen dynamic glucose monitor data through the artificial screening mode and be aided with analytic system and calculate, workload is bigger.If when the dynamic glucose monitor data that obtains increases, amount of calculation will increase greatly, greatly reduce the efficient of clinical diagnosis and scientific research and the actual effect of parameter.In addition; For guaranteeing the accuracy of MAGE numerical value; The medical personnel who screens analysis need carry out long professional training, if lacking experience of medical personnel causes existing than mistake between result of calculation and the actual value, also can reduce the accuracy and the validity of MAGE numerical value.

In order to solve the problem that MAGE numerical value calculates automatically, have reports that some calculate the MAGE numerical methods automatically, as; Peter A.Baghurst, " Calculating the Mean Amplitude of Glycemic Excursion from Continuous Glucose Monitoring Data-An Automated Algorithm ", Diabetes Technology&Therapeutics; Vol.13, pp.296-302,2011 with Gert Fritzsche; Klaus-Dieter Kohnert, Peter Heinke, Lutz Vogt; And Eckhard Salzsieder. " The use of a computer program to calculate the Mean Amplitude of Glycemic Excursions ", Diabetes Technology&Therapeutics, vol.13; Pp.319-325,2011.In the above-mentioned document, the former provides a kind of computing method based on peak value and valley point in the dynamic blood glucose level data curve of automatic identification, and this method is owing to carried out smoothing processing to dynamic blood glucose level data curve; Cause data distortion easily; The accuracy of result of calculation can be affected, and in addition, this method adopts the first order difference mode to discern knee point; Yet; Therefore the data that dynamically exist many blood glucose values to equate continuously in the blood glucose level data curve receive the influence of these data, and the first order difference mode is discerned knee point also can cause the error of calculation; The latter has mainly introduced a kind of computational analysis program of the MAGE of calculating numerical value, and the research emphasis of this article is focusing more on the feasibility and the efficiency assessment of calculating, and comparatively simple to the computing method introduction.In addition, the judgement for the effective blood glucose fluctuation in the dynamic glucose monitor data in these reports still too relies on the judgement of visual inspection and experience, therefore still has the deviation of subjectivity.

Summary of the invention

In view of there is above-mentioned deficiency in prior art; Technical matters to be solved by this invention provides a kind of computing method of MAGE; This method can be discerned the effective blood glucose fluctuation in the dynamic glucose monitor data automatically, calculates MAGE fast and accurately.

The technical scheme that the present invention addresses the above problem is following:

1. the computing method of a MAGE, this method may further comprise the steps:

1) reads in the dynamic glucose monitor data that from dynamic blood sugar monitoring instrument device, obtains, calculate the blood sugar standard deviation of the data of being obtained; Calculate the extreme point all in the data that obtains, obtain the set A of an extreme point sequence;

2) utilize differential evolution algorithm to find the solution the objective function shown in the following formula (1), meet the extreme point of following two conditions in the statistics set A, obtain set B:

Absolute value >=blood sugar the standard deviation of the blood sugar value difference that two a. adjacent extreme points are corresponding;

B. one adjacent of two extreme point are maximum value, and another is a minimal value;

$\underset{K,{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \max }{Z}_K({\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K)=\underset{K,{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \max }\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}|f\left(t_{n_{k+1}}\right)-f\left(t_{n_k}\right)|---\left(1\right)$

In the formula (1), element n ₁, n ₂..., n _kBe the element in the set A, Z _K(n ₁, n ₂..., n _k) be by element n ₁, n ₂..., n _kSubclass { the n that is formed ₁, n ₂..., n _kThe fluctuating range sum of corresponding effective blood glucose fluctuation, constant K is to satisfy the number of the extreme point of above-mentioned condition,

Be subclass { n ₁, n ₂..., n _kMiddle n _kCorresponding blood glucose value;

The absolute value of the blood sugar value difference that 3) each adjacent extreme point is corresponding among the first set of computations B calculates all average absolute again, promptly gets MAGE.

For further guaranteeing the last accuracy of calculating, the present invention also comprises the data pretreatment operation in step 1), and detailed process is:

After reading in the dynamic glucose monitor data that from dynamic blood sugar monitoring instrument device, obtains, earlier the back blood glucose value data in the deletion blood glucose value equate adjacent 2 are calculated described blood sugar standard deviation and all extreme points again.

Because the present invention adopts differential evolution algorithm to find the solution unconfined condition nonlinear programming problem objective function; Guaranteed the accurate identification of effective blood glucose fluctuation amplitude; Avoided in computation process, introducing the subjectivity factor affecting,, utilized computer aided calculation guaranteeing calculating accuracy and conforming while; Improve counting yield greatly, had higher actual effect.

Description of drawings

Fig. 1 is the process flow diagram of a specific embodiment of the computing method of a kind of MAGE according to the invention.

Fig. 2 is 24 hours dynamic change of blood sugar curve maps of a routine hospitalier.

Fig. 3 is the synoptic diagram of the data segment that blood glucose value in the dynamic glucose monitor data curve of filtering equates continuously in the computing method of a kind of MAGE according to the invention.

Fig. 4 for try to achieve in the computing method of a kind of MAGE according to the invention all effective extreme points after data and curves.

Fig. 5 is the correlation analysis result that the computing method described in the computing method of a kind of MAGE according to the invention and traditional manual calculation method are divided into groups for different pieces of information.

Fig. 6 is the computing method and the Bland-Altman evaluation graph of traditional manual calculation method for different group data described in the computing method of a kind of MAGE according to the invention.

Embodiment

Below in conjunction with accompanying drawing and embodiment the present invention is done further detailed description.

Example 1

Referring to Fig. 1, below be the computing method that example is described MAGE according to the invention in detail with 24 hours dynamic change of blood sugar curve maps of a routine hospitalier:

1) can know that from the used corresponding Monitoring Data analysis of derivation from dynamic blood sugar monitoring instrument device of hospitalier blood glucose value is the function of time variable t, can be expressed as f (t), and wherein variable t is constant duration set { t with Fig. 2 ₁, t ₂..., t _L.Visible like Fig. 2, there are many data segments that equate continuously like the blood glucose value of arrow indication among the figure in the curve, i.e. f (t _x)=f (t _X+1), wherein x representes to write down the sequence number of blood glucose value.Therefore, be to guarantee the accuracy that MAGE calculates, need to remove earlier the back blood glucose value data in adjacent 2 that blood glucose value equates, the removal process is as shown in Figure 3.And then the blood sugar standard deviation of the data after the computing; Utilize discrete series extreme point algorithm that the data after handling are asked the operation of discrete series extreme point, thus obtain the extreme point sequence of these data set 1,2 ..., N};

2) with arrangement set 1,2 ..., N} utilizes differential evolution algorithm to find the solution separating of objective function as the search volume, promptly tries to achieve all effective extreme points, concrete steps are following:

2.1) usually maximum value and the minimal value of function appearances that should interlock, therefore, when finding out all effective extreme points, the fluctuating range sum of effective blood glucose fluctuation maximum in the blood glucose level data is formulated as:

$\underset{K,{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \max }{Z}_K({\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K)=\underset{K,{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \max }\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}|f\left(t_{n_{k+1}}\right)-f\left(t_{n_k}\right)|---\left(1\right)$

In the formula (1), element n ₁, n ₂..., n _kBe set 1,2 ..., the element among the N}, Z _K(n ₁, n ₂..., n _k) be by element n ₁, n ₂..., n _kSubclass { the n that is formed ₁, n ₂..., n _kThe fluctuating range sum of corresponding effective blood glucose fluctuation, constant K is the number of effective extreme point,

Be subclass { n ₁, n ₂..., n _kMiddle n _kCorresponding blood glucose value.Subclass { n ₁, n ₂..., n _kSatisfy condition:

$\left\{\begin{array}{c}1\&le;{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1<{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2<...<n_K\&le;N\\ {(-1)}^{n_{k+1}-n_k}=-1,k=\mathrm{1,2},...,K-1\\ |f\left(t_{n_{k+1}}\right)-f\left(t_{n_k}\right)|\&GreaterEqual;\mathrm{SD},k=\mathrm{1,2},...,K-1\end{array}\right.---\left(2\right)$

In the formula (2), SD is the blood sugar standard deviation described in the step 1), and constant K is the number of effective extreme point, 2≤K≤N.

Convolution (1) and formula (2) can know, it is that constraint condition is the finding the solution of nonlinear programming problem of objective function to formula (1) that the calculating of MAGE is with formula (2), if there is a feasible solution in this problem for constant K

With a maximal value

So, also necessarily there is a maximal value corresponding to constant K-1

And If K ^*Be the maximal value of the feasible solution constant K of this problem, then feasible solution Be MAGE and calculate all required effective extreme point arrangement sets.

2.2) utilize differential evolution algorithm to find the solution through to solution procedure 2.1) gained has the nonlinear programming problem objective function of constraint condition, thereby obtain feasible solution

concrete steps be:

2.2.1) utilize the penalty algorithm, make formula (3) be penalty and initialization penalty coefficient, convolution (2) is converted into formula (1) the nonlinear programming problem objective function of unconfined condition:

$\left\{\begin{array}{c}{g}_{k({\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K)}-\mathrm{SD}-|f_{n_{k+1}}-f_{n_k}|\\ h_{k({\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K)}={(-1)}^{n_{K+1}-n_k}+1,k=\mathrm{1,2},...,K-1\end{array}\right.---\left(3\right)$

$\underset{{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \min }{Y}_K(n_1,n_2,...,n_K)$

$=\underset{K,{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA00002104964100011.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA00002104964100021.png,HDA00002104964100031.png"\; state="\{\{state\}\}">n</figure-callout>}}_2,...,n_K}{\arg \min }[-\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}|f_{n_{k+1}}-f_{n_k}|+\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}{\mathrm{\&mu;}}_k\max {\{0,g_k\}}^2+\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_k{h}_k^2]---\left(4\right)$

In the formula (4), μ _kAnd λ _kBe penalty coefficient, and all trend towards+∞, can know minimum value Y by formula (4) when formula (4) _K ^*＞0 o'clock, the nonlinear programming problem objective function of this unconfined condition did not have feasible solution.

2.2.2) utilize differential evolution algorithm solution procedure 2.2.1) gained unconfined condition nonlinear programming problem objective function:

2.2.2A) related parameter in the initialization differential evolution algorithm: comprise population quantity NP, maximum iteration time G _MaxAnd iteration count G=1, solution space largest N, crossover probability CR;

2.2.2B) create an initialization population and this population is uniformly distributed in whole search volume of separating at random, the representation of population as shown in the formula:

$N_i^0=(\mathrm{1,1},...,1)+\mathrm{NINT}[({\mathrm{\&rho;}}_{i1},{\mathrm{\&rho;}}_{i2},...,{\mathrm{\&rho;}}_{\mathrm{iN}})\&CenterDot;(N-1)]---\left(5\right)$

In the formula (5), i is the index of current trial vector, Be object vector, ρ _I1Be the random number between [0,1], NINT [B] expression is done the round operation to vectorial B;

2.2.2C) begin from K=2 dimension search volume, search for whole search volume, equal the set sizes of extreme point until the search volume of feasible solution size, concrete steps are:

2.2.2D) feasible solution of ferret out function in working as former generation, for each object vector for G In the same generation, randomly draw other three different individual vectors

With

R wherein ₂, r ₂, r ₂Be in [1, NP] interval, randomly draw, mutual unequal three integers, and all unequal with the index i of current trial vector.Obtain a corresponding with it variation vector

through the difference strategy so

$V_i^G=N_{r_1}^G+\mathrm{NINT}[F\&times;(N_{r_2}^G-N_{r_3}^G)]---\left(6\right)$

In the formula (6), F is the difference vector zoom factor of one [0,1] interval interior value;

2.2.2E) in step 2.2.2D) the resulting variation vector

and the corresponding target vector

between the crossover operation to generate a test vector

$u_{\mathrm{ji}}^G=\left\{\begin{array}{cc}{v}_{\mathrm{ji}}^G,& \mathrm{rand}\left[\mathrm{0,1}\right]\&le;\mathrm{CRorj}=j_{\mathrm{rand}},\\ n_{\mathrm{ji}}^G,& \mathrm{otherwise},\end{array}\right.j=\mathrm{1,2},...,n.---\left(7\right)$

CR ∈ [0,1] is a crossover probability that is set by differential evolution method user in the formula (7), j _RandBe the index of in [1, NP] interval, selecting at random, this index is used for guaranteeing

Can

In obtain a parameter value at least.The numeric parameter of object vector is according to arranging in proper order from small to large;

2.2.2F) adopt greedy search strategy to step 2.2.2E) resulting trial vector

compares selection with object vector

; If the functional value

that trial vector

is determined is replaced by the trial vector when former generation less than the then follow-on object vector of functional value

that object vector is determined; As denying; Then the object vector with current band remains to the next generation, with formula (8) expression as follows:

$N_i^{G+1}=\left\{\begin{array}{c}{U}_i^G,\mathrm{ifY}\left(U_i^G\right)<Y\left(N_i^G\right),\\ N_i^G,\mathrm{otherwise}.\end{array}\right.---\left(8\right);$

2.2.2G) make iteration count increase progressively 1, if current iteration counter values G is less than maximum iteration time G _Max, repeating step 2.2.2D then), 2.2.2E), 2.2.2F), in the next generation, search for feasible solution, if current iteration counter values G equals maximum iteration time G _Max, then jump out iteration and calculate the maximal value of the fluctuating range sum of all effective blood glucose fluctuations

If

Be not more than 0, then make the solution space dimension increase progressively 1, and repeating step 2.2.2D), 2.2.2E), 2.2.2F), if

Greater than 0, then differential evolution calculates and finishes, this moment Be the maximal value of constraint condition nonlinear programming problem objective function, the corresponding feasible solution that is constraint condition nonlinear programming problem objective function of separating

Promptly try to achieve all effective extreme point arrangement sets, as shown in Figure 4;

3) according to step 2.2.2G) all effective extreme point arrangement sets, calculate MAGE, computation process is following:

A) if

corresponding be first local minizing point of dynamic glucose monitor data, then the computing formula of MAGE is:

B) if

corresponding be first local maximum point of dynamic glucose monitor data, then the computing formula of MAGE is:

In formula (9) and the formula (10), MAGE ₊For the direction (from the trough to the crest or from the crest to the trough) that effectively fluctuates with first is the MAGE of calculated direction, MAGE _-For the opposite side (from the crest to the trough or from the trough to the crest) with this calculated direction is the MAGE of calculated direction, MAGE _aBe MAGE ₊And MAGE _-Mean value.

Example 2 (accuracy of method and and feasibility checking)

Following table 1 is depicted as the MAGE that utilizes traditional manual calculation method and computing method of the present invention to calculate respectively from 76 routine hospitaliers' dynamic glucose monitor data, is divided into 3 groups, wherein; Normal adult 27 examples; Diabetes B adult patient 25 examples, normal pregnancies 24 examples, MAGE _OExpression adopts traditional manual calculation method to calculate the averaging of income blood glucose fluctuation width of cloth, MAGE _CExpression adopts computing method of the present invention to calculate averaging of income blood glucose fluctuation amplitude.Owing to adopt the accuracy of traditional manual calculation method gained result of calculation influenced by medical personnel's subjectivity, so MAGE in this table _ONumerical value is the staff who is engaged in related work for a long time and adopts traditional manual calculation method to obtain, and has higher accuracy.

3 groups computing method gained result of calculation of the present invention and traditional manual calculation method gained result of calculation in the table 1 is carried out correlation analysis, analysis result is as shown in Figure 5, the related coefficient γ of normal adult group=0.994 (P＜0.01) at present; Referring to Fig. 5 a), related coefficient γ=0.997 of diabetes B adult patient group (P＜0.01) is referring to Fig. 5 b); The related coefficient γ of normal pregnancies group=0.998 (P＜0.01); Referring to Fig. 5 c), 76 routine hospitaliers' related coefficient γ=0.997 (P＜0.01) is referring to Fig. 5 d); Hence one can see that, computing method gained result of calculation MAGE of the present invention _CWith traditional manual calculation method gained result of calculation MAGE _OFor the different hospitaliers equal significant correlation that divides into groups.

Pass through the consistance of Bland-Altman graph evaluation computing method of the present invention and traditional manual calculation method again; The gained result is as shown in Figure 6; Different hospitaliers adopt the difference of two kinds of distinct methods gained result of calculations to be respectively; Normal adult group (0.00242 ± 0.0136) mmol*L-1; Diabetes B adult patient group (0.01533 ± 0.0275) mmol*L-1, normal pregnancies group (0.0034 ± 0.0191) mmol*L-1,76 routine hospitalier (0.00526 ± 0.0218) mmol*L-1.95% consistance boundary utilizes mean ± 1.96SD formula to try to achieve, and is respectively normal adult group (0.0243 ,-0.0292); Diabetes B adult patient group (0.0692 ,-0.0386); Normal pregnancies group (0.0409 ,-0.0341); 76 routine hospitaliers (0.0480 ,-0.0374).Bland-Altman figure different in the analysis chart 6 can know, drop on the outer point of consistance boundary and are respectively 3 of normal adult groups, account for 11.1% (3/27); 1 of diabetes B adult patient group accounts for 4% (1/25); 1 of normal pregnancies group accounts for 4.2% (1/24); 6 of 76 routine hospitaliers account for 7.9% (6/76).Analysis result shows that MAGE computing method of the present invention (MAGEc) and traditional manual calculation method (MAGEo) have consistance preferably.

Table 1 computing method according to the invention and the contrast of traditional manual calculation method data

Claims (2)

1. the computing method of a MAGE, this method may further comprise the steps:

1) reads in the dynamic glucose monitor data that from dynamic blood sugar monitoring instrument device, obtains, calculate the blood sugar standard deviation of the data of being obtained; Calculate the extreme point all in the data that obtains, obtain the set A of an extreme point sequence;

2) utilize differential evolution algorithm to find the solution the objective function shown in the following formula (1), meet the extreme point of following two conditions in the statistics set A, obtain set B:

Absolute value >=blood sugar the standard deviation of the blood sugar value difference that two a. adjacent extreme points are corresponding;

B. one adjacent of two extreme point are maximum value, and another is a minimal value;

$\underset{K,n_1,n_2,...,n_K}{\arg \max }{Z}_K(n_1,n_2,...,n_K)=\underset{K,n_1,n_2,...,n_K}{\arg \max }\underset{k=1}{\overset{K-1}{\mathrm{\&Sigma;}}}|f\left(t_{n_{k+1}}\right)-f\left(t_{n_k}\right)|---\left(1\right)$

In the formula (1), element n ₁, n ₂..., n _kBe the element in the set A, Z _K(n ₁, n ₂..., n _k) be by element n ₁, n ₂..., n _kSubclass { the n that is formed ₁, n ₂..., n _kThe fluctuating range sum of corresponding effective blood glucose fluctuation, constant K is to satisfy the number of the extreme point of above-mentioned condition,

Be subclass { n ₁, n ₂..., n _kMiddle n _kCorresponding blood glucose value;

The absolute value of the blood sugar value difference that 3) each adjacent extreme point is corresponding among the first set of computations B calculates all average absolute again, promptly gets MAGE.

2. the computing method of a kind of MAGE according to claim 1 is characterized in that, wherein step 1) also comprises following data pre-treatment step:

After reading in the dynamic glucose monitor data that from dynamic blood sugar monitoring instrument device, obtains, earlier the back blood glucose value data in the deletion blood glucose value equate adjacent 2 are calculated described blood sugar standard deviation and all extreme points again.

Images