CN107436993A - Establish the method and server of ICU conditions of patients assessment models - Google Patents
Establish the method and server of ICU conditions of patients assessment models Download PDFInfo
- Publication number
- CN107436993A CN107436993A CN201710310478.8A CN201710310478A CN107436993A CN 107436993 A CN107436993 A CN 107436993A CN 201710310478 A CN201710310478 A CN 201710310478A CN 107436993 A CN107436993 A CN 107436993A
- Authority
- CN
- China
- Prior art keywords
- data
- icu
- vital sign
- characteristic vector
- analysis model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
This application provides a kind of method and server for establishing ICU conditions of patients assessment models, this method includes：Obtain the vital sign data and treatment results of ICU historic patients；Feature extraction is carried out to vital sign data and obtains corresponding characteristic vector, according to the analysis model between characteristic vector training feature vector and treatment results；Analysis model is used for the treatment results of the current patents according to the vital signs in real time data assessment of ICU current patents.Its treatment results is assessed using the vital sign data of the analysis model and combination ICU current patents, so as to which treatment results are produced with reliable prediction.
Description
Technical field
The application is related to Analysis of Medical Treatment Data field, more particularly to a kind of method for establishing ICU conditions of patients assessment models
And server.
Background technology
Hospital intensive care room (Intensive Care Unit, ICU) is that critical illness is intensive, the state of an illness is changeable, crisis clump
Raw place, place continuous or close to continuous observation, diagnosis and treatment and monitoring is carried out to urgent patient.ICU treatments are slight not
The risk of patient death will be greatly increased in time.Producing these early stages for jeopardizing situation, some uncommon vital signs
Just occur, and along with common complication.Current medical ICU treatment means rely primarily on the experience and association area of doctor
Accumulation of knowledge so that operating efficiency and diagnosis and treatment are of low quality.At present evaluation areas, various medical conditions are monitored in conditions of patients
Under have many points-scoring systems using medical knowledge.For example, the result in patient's kidney failure can use Acute Physiology Score (12 lifes
Manage variable), chronic healthy scoring (organ dysfunction) is predicted, and is scored using APACHE II and carried out total evaluation.This
A little traditional medical science methods of marking can be assessed the state of an illness of patient to a certain extent, but score-system is mechanical, individual character
Change degree is low, and effectively support can not be produced to the further decision-making of doctor.
In recent years, developing rapidly with electronic information technology, hospital information system (Hospital Information
System, HIS) and digital Medical Devices extensive use, medical treatment constantly expands with health data amount, and database and distribution are literary
The development of the technologies such as part system solves the problems, such as the storage of mass data and data effectiveness of retrieval, but can not change that " data are quick-fried
Fried but knowledge is poor " phenomenon.ICU is the place that much information crosses in clinical department, a large amount of patient's bodies of the scene set
Levy data, including the various vital sign parameters such as blood pressure, body temperature, heart rate, electrocardio；Great amount of images continuous monitoring data, including sound
Video image information or even ultrasonic portable set monitoring information；Large number quipments Monitoring Data, including advanced physiological monitor, rise
Fight defibrillation monitor, cardio-pulmonary resuscitation machine, it is advanced transhipment ventilator data；A large amount of patient medical record data, such as electronic health record
Data interaction all be present in systems such as (Electronic Medical Record, EMR).But number comprehensive to ICU at present, continuous
Effectively utilized according to lacking.
The content of the invention
In view of this, the embodiment of the present application provides a kind of method and server for establishing ICU conditions of patients assessment models,
To solve the utilization of ICU medical datas in the prior art and excavate insufficient, the relatively low technical problem of the level of informatization.
According to the one side of the embodiment of the present application, there is provided a kind of method for establishing ICU conditions of patients assessment models, institute
The method of stating includes：Obtain the vital sign data and treatment results of ICU historic patients；Feature is carried out to the vital sign data
Extraction obtains corresponding characteristic vector, and the analysis mould between the characteristic vector and treatment results is trained according to the characteristic vector
Type；The analysis model is used for the treatment knot of the current patents according to the vital signs in real time data assessment of ICU current patents
Fruit.
According to the another aspect of the embodiment of the present application, there is provided a kind of server includes：Processor；For storing processor
The memory of executable instruction；Wherein, the processor is configured as：Obtain the vital sign data of ICU historic patients and control
Treat result；Feature extraction is carried out to the vital sign data and obtains corresponding characteristic vector, is trained according to the characteristic vector
Analysis model between the characteristic vector and treatment results；The analysis model is used for the real-time life according to ICU current patents
Order the treatment results that sign data assesses the current patents.
The beneficial effect of the embodiment of the present application includes：Vital sign number is trained using the medical data of ICU historic patients
According to the analysis model between treatment results, the analysis model and the vital sign number of combination ICU current patents are further utilized
According to its treatment results is assessed, so as to which treatment results are produced with reliable prediction.
Brief description of the drawings
By the description to the embodiment of the present application referring to the drawings, the above-mentioned and other purpose of the application, feature and
Advantage will be apparent from, in the accompanying drawings：
Fig. 1 is the schematic flow sheet for the method that the embodiment of the present application establishes ICU conditions of patients assessment models；
Fig. 2 is the schematic flow sheet for the method that the embodiment of the present application establishes ICU conditions of patients assessment models；
Fig. 3 is the schematic flow sheet for the method that the embodiment of the present application establishes ICU conditions of patients assessment models.
Embodiment
The application is described below based on embodiment, but the application is not restricted to these embodiments.Under
Text is detailed to describe some specific detail sections in the detailed description of the application.Do not have for a person skilled in the art
The description of these detail sections can also understand the application completely.In order to avoid obscuring the essence of the application, known method, mistake
The not narration in detail of journey, flow, element and circuit.
In addition, it should be understood by one skilled in the art that provided herein accompanying drawing be provided to explanation purpose, and
What accompanying drawing was not necessarily drawn to scale.
Unless the context clearly requires otherwise, otherwise entire disclosure is similar with the " comprising " in claims, "comprising" etc.
Word should be construed to the implication included rather than exclusive or exhaustive implication；That is, it is containing for " including but is not limited to "
Justice.
In the description of the present application, it is to be understood that term " first ", " second " etc. are only used for describing purpose, without
It is understood that to indicate or implying relative importance.In addition, in the description of the present application, unless otherwise indicated, the implication of " multiple "
It is two or more.
The embodiment of the present application to the medical data of ICU historic patients fully excavate and analyze, and trains an analysis
Model, the therapeutic effect of ICU current patents is assessed using the analysis model, to improve the success rate of ICU treatments.This point
Analysis model can automatically update with the accumulation of ICU medical datas, make assessment result more and more accurate, moreover it is possible to what is got
The vital sign data of ICU current patents is adjusted and assessed, and success rate highest and cost are searched out with assist personnel
Minimum medical scheme.
Fig. 1 is a kind of method for establishing ICU conditions of patients assessment models that the embodiment of the present application provides, suitable for service
Device, this method comprise the following steps.
S10, obtain the vital sign data and treatment results of ICU historic patients.
ICU historic patients refer to the patient for receiving ICU diagnosis and treatment in the court, including have been discharged from hospitals upon recovery and dead trouble
Person, the sample data of the medical datas of these historic patients as training analysis model.
Vital sign data reflect patient's items physical signs, including but not limited to personal information (such as the age, sex,
Height, body weight etc.), checking information (such as blood routine, routine urinalysis, stress etc.), image information (such as CT, nuclear magnetic resonance,
Ultrasound etc.), medical information (such as treatment, medical personnel's operation, medication etc.), custodial care facility during electronic health record information and ICU
The various forms such as timing information (such as ECG monitor, lung ventilator, blood pressure etc.) mass data.
Treatment results are a kind of label datas, reflect treatment effect of the patient in ICU, such as represent and " take a turn for the better, deteriorate, be dead
Die " etc. information label data.
The sample of the vital sign data and treatment results of ICU historic patients as training analysis model is obtained from database
Notebook data.
S11, feature extraction is carried out to vital sign data and obtains corresponding characteristic vector, according to characteristic vector training characteristics
Analysis model between vector and treatment results；Analysis model is used to be commented according to the vital signs in real time data of ICU current patents
Estimate the treatment results of current patents.
Vital sign data is analyzed using feature extraction algorithm, obtains its corresponding characteristic vector.Further according to
This feature is vectorial and its corresponding treatment results training analysis model.
In one embodiment of the application, the analysis model can be obtained by the following method.Assuming that the number of a patient
According to including：
Xi=[x1,x2,x3,...xn]；
y；
Wherein, XiRepresent the vital sign data collection of the patient, x1,x2,x3,...xnCorresponding Patient height, body weight, blood pressure,
The vital sign data of each dimension such as blood oxygen, share n dimension；Y represents the treatment results label data of patient, corresponding treatment
As a result one kind in (such as improvement, deterioration, death etc.).
Accordingly, the data set that substantial amounts of historic patient data is formed is：
X=[X1,X2,X3,...Xn]T；
Y=[y1,y2,y3,...ym]T。
Data processing and feature extraction are carried out for different types of data in vital sign data, by different types of number
According to extract feature composition feature vector, X '=ffeature(X), wherein, ffeatureFunction is data processing and feature extraction letter
Number, effect is that the feature set X' for being suitable for training analysis model is processed into from original X data pooled applications respective algorithms.
When carrying out data processing, different data processing functions is used for different data types.For example, when being directed to
Ordinal number evidence and view data, interpolative operation can be done to missing values, the processing function such as smoothing processing is done to exceptional value；For discrete
Data, missing values and exceptional value can be replaced with the average of same type, intermediate value or other statistics.Carrying out feature extraction
When, equally use different feature extraction functions for different data types.Such as time series data, when retaining one section
Between time series data, then extract the information of time domain and frequency domain；The extraction of temporal signatures includes but is not limited to calculate this section of sequential
The statistical parameters such as the averages of data, variance, each rank norm, various series expansions, the extraction of frequency domain character include but is not limited to profit
The various spectrum signature information extracted with mathematical tools such as Fourier transformation, wavelet transformations.For other kinds of data, then may be used
Using common feature extraction function corresponding with its data type, can be gone out when necessary using multiple attributes extractions one or
Multiple features.
Utilize characteristic vector corresponding to vital sign data and treatment results label data, training analysis model：
Y'=Fθ(X')；
F=arg min ∑s (log (Y')-log (Y))2；
Wherein, FθRepresentative model function, Y' are the judged results that analysis model trains to obtain according to history data set；F is Fθ
Constraints (i.e. choose and training analysis model constraints)；θ is to make error function ∑ (log (Y')-log (Y))2Most
Small vector value, belong to the vector of the error function solution space., i.e., assess what is obtained for all patients, above-mentioned analysis model
Overall error in treatment results Y' and historical data between the actual treatment results Y of patient is minimum.It is continuous with historical data
Accumulation, the assessment result of the analysis model is more accurate, and overall error is smaller.
In the present embodiment, trained using the medical data of ICU historic patients between vital sign data and treatment results
Analysis model, using the analysis model and combine ICU current patents vital sign data assess its treatment results, so as to
Treatment results are produced with reliable prediction, using the above-mentioned method for establishing analysis model, the less analysis of error can be trained
Model, and as the data accumulation of historic patient, the continuous renewal of analysis model, resultant error also can be smaller.Below to more
The process of new analysis model is described further, and in one embodiment, this method further comprises the steps.
S12, according to the vital sign data of the ICU historic patients accumulated in the last statistics duration and treatment results more
New analysis model.
Utilize the vital sign data and treatment results replacement analysis model of accumulation：
Y'=Fθ(X')；
F=arg min ∑s (log (Y')-log (Y))2。
To wherein F=arg min ∑s (log (Y')-log (Y))2It is rewritten as：
J (θ)=∑ (log (F (X'))-log (Y))；
minθJ(θ)。
Local derviation J is sought for function J (θ)：
Wherein, θiThe value before renewal is represented,The amount by gradient direction reduction is represented, α represents step-length, also
It is the variable quantity every time on the direction of gradient reduction.
For vectorial θ, per one-dimensional component θiThe direction of a gradient can be obtained, so as to can determine that an entirety side
To when change being changed can towards the most direction of decrement reaches a smallest point, and no matter it is local
Or global, then F corresponding to this smallest pointθIt is exactly the minimum disaggregated model F of root-mean-square error.
In the present embodiment, towards the side that decrement on each gradient direction of error function is most when being updated to analysis model
To being changed, make its root-mean-square error result as small as possible, so that the error of analysis model is also smaller therewith.
After training obtains analysis model, you can the treatment results of ICU current patents are predicted using analysis model.
In one embodiment, as shown in Fig. 2 this method further comprises the steps.
S13, obtain the vital signs in real time data of ICU current patents and carry out feature extraction, obtain vital signs in real time
Characteristic vector corresponding to data.
ICU current patents refer to the patient currently just guarded in ICU.For ICU current patents, by current ICU diseases
Vital signs in real time data caused by room medical personnel and custodial care facility with step S11 identicals data processing and feature extraction
Algorithm, construct the characteristic vector of description current patents' vital sign.
The data processing and feature extraction of vital signs in real time data can be set according to the sampling interval of ICU custodial care facilities
In the counting period, new data are sampled every time one data processing and feature extraction computing are carried out to time series data.For other
Data type, it may be reused after once calculating.When handling real-time time series data, system retains a timing
Long historical data, it is not lost with the information ensured in time series data.Finally, combinations of features different type extracted exists
Together, characteristic vector corresponding to the vital signs in real time data of current patents is formed.
S14, characteristic vector corresponding to vital signs in real time data is inputted into analysis model, assesses the treatment knot of current patents
Fruit.
For example, the vital signs in real time data of current patents are Xnew=[x1,x2,x3,...xn], using data processing and
Obtained after feature extraction algorithm feature vector, X corresponding to the vital signs in real time data of the current patents 'new, utilize analysis model
Assessment obtains the treatment results Y' of the current patentsnew=F (X'new)。
Y'newThe treatment results for the current patents that representative is assessed to obtain using analysis model, can represent " take a turn for the better, deteriorate
And death " etc. one kind in the label data of result.The different treatment results obtained based on assessment, if showing the life of patient
Sign situation can will send early warning letter than further worsened at present (danger that deterioration or death be present), system to medical personnel
Breath, to prompt it to improve therapeutic scheme or take emergency measures in advance.
In the present embodiment, the treatment results of ICU current patents are predicted using analysis model, contribute to medical personnel
The danger that sb.'s illness took a turn for the worse is found in time, to change therapeutic scheme as early as possible, improves the survival rate of ICU patient, and predicting
During the result of deterioration or death, early warning can be sent to medical personnel, to take counter-measure in time.In addition, another
In individual embodiment, successive treatment scheme can also be determined using the analysis model assist personnel automatically, as shown in figure 3, should
Method further comprises the steps.
S15, the vital signs in real time data got are adjusted automatically, by the vital signs in real time data after adjustment
Corresponding characteristic vector inputs analysis model, and the treatment results for obtaining to improve after assessing can also make the Least-cost of adjustment
Vital sign data Adjusted Option.
Medical personnel and custodial care facility gather vital sign data in real time in ICU wards, and these data include heart rate, blood
The special index of the medical treatment such as oxygen, blood pressure, they medically have a normal value section.With reference to the current vital sign number of patient
According to the medically index normal value section, it can be determined which Indexes Abnormality of patient and abnormal degree gone out.Known according to medical science
Know, the cost that controlling different indexs needs to pay also differs.For example, so uncontrollable index of doctor such as sex, age,
Cost is infinity；And the controllable index such as body temperature, blood pressure, blood oxygen, then can be according to the difficulty or ease journey for controlling these indexs
Spend to define cost.After the treatment results of patient are assessed every time using analysis model, system can Binding change difference index hardly possible
The cost matrix of easy degree chooses and change automatically one or more and has produced abnormal index, and these abnormal indexes are adjusted
It is whole arrive normal value section after, be re-entered into analysis model and the treatment results of the patient assessed.By being commented before contrast
The treatment results estimated, to determine the treatment results for controlling these indexs whether can be effectively improved patient, and combine and control different fingers
Target cost, point out which kind of sequentially changes which kind of index can at utmost change while minimum cost is paid by medical personnel
The treatment results of kind patient, that is, provide more rational diagnosis and treatment suggestion in next step.
The cost matrix for changing vital sign is C=[C1,C2,C3,...Cn]；The current vital sign data of patient is
Xnew=[x1,x2,x3,...xn].Using analysis model explore the optimal diagnosis and treatment suggestion of patient method be：If i, j belong to [1, n]
Section, xjFor the data of some dimension in the current vital sign data of patient, xiFor what is be adjusted to required for the dimension of searching
Target data.Vital sign data X after adjustmenti=[x1,x2,x3,...xj→i,...xn], wherein xj→iRepresent the dimension data
From xjIt is adjusted to xi。
Vital sign data X after being adjusted according to feature extraction algorithmiCorresponding feature vector, X 'i, by characteristic vector
X'iInput analysis model obtains corresponding assessment result Y'i=F (X'i)；If obtained assessment result Y'iMeet simultaneously following
Condition：
(1)Y'iBetter than Y'new, wherein Y'newRepresent the vital signs in real time data X of current patentsnewResulting assessment
As a result；
(2)arg min(X'iC), i.e., feature vector, X 'iTotal Least-cost of corresponding data point reuse；
When above-mentioned condition (1) and (2) while when meeting, then by the dimension data from xjIt is adjusted to xi。
In the present embodiment, vital sign data and utilization analysis model that adjust automatically is got assess the treatment after adjustment
As a result and its corresponding cost, the diagnosis and treatment suggestion that treatment results and can pays minimum cost can be improved by finding, to aid in medical care people
The dicision of diagnosis and treatment of member's next step is more reasonable, effective.
In system architecture, the embodiment of the present application can independently set up a set of server apparatus and carry out training analysis model, pre-
Survey treatment results and provide diagnosis and treatment suggestion, or directly with hospital's ICU existing equipments and the system integration, according to real time data from
It is dynamic directly to produce assessment result, to find the potential danger of patient earlier.On the basis of potential danger is found earlier, more enter
One step, can be according to the vital sign index of assessment result integrative medicine professional knowledge adjust automatically patient, and combine index
The cost matrix of adjustment, provide the diagnosis and treatment suggestion of science.
In addition, in the embodiment of the present application, server can be by hardware processor (hardware processor) come real
Existing above-mentioned each functional steps.Server includes：Processor, for storing the memory of processor-executable instruction；Wherein, locate
Reason device is configured as：Obtain the vital sign data and treatment results of ICU historic patients；Feature is carried out to vital sign data to carry
Corresponding characteristic vector is obtained, according to the analysis model between characteristic vector training feature vector and treatment results；Analyze mould
Type is used for the treatment results of the vital signs in real time data assessment current patents according to ICU current patents.
In one embodiment, feature extraction is carried out to vital sign data and obtains corresponding characteristic vector, according to feature
Analysis model between vectorial training feature vector and treatment results includes：For different types of data in vital sign data
Feature extraction is carried out, generates characteristic vector corresponding to vital sign data；Analysis mould is established according to characteristic vector and treatment results
Type, i.e.,
Y'=Fθ(X')；
F=arg min ∑s (log (Y')-log (Y))2；
Wherein, X' representative features vector, FθRepresentative model function, Y' represent analysis model and train to obtain according to characteristic vector
Assessment result；F is FθConstraints, θ is to make error function ∑ (log (Y')-log (Y))2Minimum vector value, i.e., for
It is minimum that all patients make analysis model assess the overall error between the treatment results Y of obtained Y' and historic patient.
In one embodiment, the processor is configured to：According to the ICU accumulated in the last statistics duration
The vital sign data and treatment results replacement analysis model of historic patient.
In one embodiment, according to the vital sign data of the ICU historic patients accumulated in the last statistics duration
Include with treatment results replacement analysis model：
According to characteristic vector corresponding to the renewal of the vital sign data of the ICU historic patients of accumulation, while cause after updating
The obtained assessment result of characteristic vector and original assessment result between root-mean-square error it is minimum.
In one embodiment, the processor is configured to：Obtain the vital signs in real time number of ICU current patents
According to and carry out feature extraction, obtain characteristic vector corresponding to vital signs in real time data；By corresponding to vital signs in real time data
Characteristic vector inputs analysis model, assesses the treatment results of current patents.
In one embodiment, the processor is configured to：If the treatment results for assessing current patents are weaker than
The real-time status of current patents, then send warning information.
In one embodiment, the processor is configured to：Automatically the vital signs in real time data to getting
It is adjusted, characteristic vector corresponding to the vital signs in real time data after adjustment is inputted into analysis model, obtains improving and comments
Treatment results after estimating can also make the vital sign data Adjusted Option of the Least-cost of adjustment.
In one embodiment, the vital signs in real time data got are adjusted automatically, will be real-time after adjustment
Characteristic vector corresponding to vital sign data inputs analysis model, and the treatment results for obtaining to improve after assessing can also make adjustment
The vital sign data Adjusted Option of Least-cost include：
Cost matrix C=[C are established according to the adjustment cost of each dimension data in vital sign data1,C2,C3,
...Cn]；Wherein C1…CnRepresent the adjustment cost of each dimension data；
In the vital sign data X of current patentsnew=[x1,x2,x3,...xn] in, it is some dimension data xjFound
Adjustment after target data xi, the vital sign data X after being adjustedi=[x1,x2,x3,...xj→i,...xn], wherein
xj→iThe dimension data is represented from xjIt is adjusted to xi；
Vital sign data X after being adjusted according to feature extraction algorithmiCorresponding feature vector, X 'i, by characteristic vector
X'iInput analysis model obtains corresponding assessment result Y'i；
If obtained assessment result Y'iMeet following condition simultaneously：
Y'iBetter than Y'new, wherein Y'newRepresent the vital signs in real time data X of current patentsnewResulting assessment knot
Fruit；
arg min(X'iC), i.e., feature vector, X 'iTotal Least-cost of corresponding data point reuse；
Then by the dimension data from xjIt is adjusted to xi。
In one embodiment, vital sign data includes personal information, checking information, image information, medical information, electricity
The combination of much information in the timing information of sub- medical record information and ICU custodial care facility.
It will be understood by those skilled in the art that embodiments herein can be provided as method, apparatus (equipment) or computer
Program product.Therefore, in terms of the application can use complete hardware embodiment, complete software embodiment or combine software and hardware
Embodiment form.Moreover, the application can use the meter for wherein including computer usable program code in one or more
The computer journey that calculation machine usable storage medium is implemented on (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of sequence product.
The application is the flow chart with reference to method, apparatus (equipment) and computer program product according to the embodiment of the present application
And/or block diagram describes.It should be understood that can be by each flow in computer program instructions implementation process figure and/or block diagram
And/or square frame and the flow in flow chart and/or block diagram and/or the combination of square frame.These computer programs can be provided to refer to
The processors of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is made to produce
One machine so that produced by the instruction of computer or the computing device of other programmable data processing devices for realizing
The device for the function of being specified in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory, which produces, to be included referring to
Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one square frame of block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that counted
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented processing, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
The preferred embodiment of the application is the foregoing is only, is not limited to the application, for those skilled in the art
For, the application can have various changes and change.All any modifications made within spirit herein and principle, it is equal
Replace, improve etc., it should be included within the protection domain of the application.
Claims (10)
- A kind of 1. method for establishing ICU conditions of patients assessment models, it is characterised in that methods described includes：Obtain the vital sign data and treatment results of ICU historic patients；Feature extraction is carried out to the vital sign data and obtains corresponding characteristic vector, according to characteristic vector training Analysis model between characteristic vector and treatment results；The analysis model is used for the real-time vital body according to ICU current patents Levy the treatment results of current patents described in data assessment.
- 2. according to the method for claim 1, it is characterised in that feature extraction is carried out to the vital sign data and obtained pair The characteristic vector answered, the analysis model between the characteristic vector and treatment results is trained to include according to the characteristic vector：Feature extraction is carried out for different types of data in the vital sign data, it is corresponding to generate the vital sign data Characteristic vector；The analysis model is established according to the characteristic vector and treatment results, i.e.,Y'=Fθ(X')；F=arg min ∑s (log (Y')-log (Y))2；Wherein, X' represents the characteristic vector, FθRepresentative model function, Y' represent the analysis model and trained according to characteristic vector Obtained assessment result；F is FθConstraints, θ is to make error function ∑ (log (Y')-log (Y))2Minimum vector value, i.e., The analysis model is set to assess the overall error between the treatment results Y of obtained Y' and historic patient for all patients minimum.
- 3. according to the method for claim 1, it is characterised in that methods described also includes：Divide according to the vital sign data of the ICU historic patients accumulated in the last statistics duration and treatment results renewal Analyse model.
- 4. according to the method for claim 3, it is characterised in that according to the ICU history accumulated in the last statistics duration The vital sign data and treatment results of patient, which updates the analysis model, to be included：According to characteristic vector corresponding to the renewal of the vital sign data of the ICU historic patients of the accumulation, at the same cause it is described more Root-mean-square error between assessment result and original assessment result that characteristic vector after new obtains is minimum.
- 5. the method according to claim 1 or 3, it is characterised in that methods described also includes：Obtain the vital signs in real time data of ICU current patents and carry out feature extraction, obtain the vital signs in real time data Corresponding characteristic vector；Characteristic vector corresponding to the vital signs in real time data is inputted into the analysis model, assesses controlling for the current patents Treat result.
- 6. according to the method for claim 5, it is characterised in that methods described also includes：If assess the current patents Treatment results be weaker than the real-time status of the current patents, then send warning information.
- 7. according to the method for claim 5, it is characterised in that methods described also includes：Automatically the vital signs in real time data got are adjusted, the vital signs in real time data after adjustment are corresponding Characteristic vector input the analysis model, the treatment results for obtaining to improve after assessing can also make the cost of the adjustment most Small vital sign data Adjusted Option.
- 8. according to the method for claim 7, it is characterised in that the vital signs in real time number to getting automatically According to being adjusted, characteristic vector corresponding to the vital signs in real time data after adjustment is inputted into the analysis model, obtaining can Improving the treatment results after assessing can also include the vital sign data Adjusted Option of the Least-cost of the adjustment：Cost matrix C=[C are established according to the adjustment cost of each dimension data in the vital sign data1,C2,C3, ...Cn]；Wherein C1…CnRepresent the adjustment cost of each dimension data；In the vital sign data X of current patentsnew=[x1,x2,x3,...xn] in, it is some dimension data xjThe tune found Target data x after wholei, the vital sign data X after being adjustedi=[x1,x2,x3,...xj→i,...xn], wherein xj→iGeneration The table dimension data is from xjIt is adjusted to xi；Vital sign data X after being adjusted according to feature extraction algorithmiCorresponding feature vector, X 'i, by feature vector, X 'i Input the analysis model and obtain corresponding assessment result Y'i；If obtained assessment result Y'iMeet following condition simultaneously：Y'iBetter than Y'new, wherein Y'newRepresent the vital signs in real time data X of current patentsnewResulting assessment result；arg min(X'iC), i.e., feature vector, X 'iTotal Least-cost of corresponding data point reuse；Then by the dimension data from xjIt is adjusted to xi。
- 9. according to the method for claim 1, it is characterised in that the vital sign data includes personal information, examines letter Breath, image information, medical information, electronic health record information and ICU custodial care facilities timing information in much information combination.
- A kind of 10. server, it is characterised in that including：Processor；For storing the memory of processor-executable instruction；Wherein, the processor is configured as：Perform claim requires to establish the assessment of ICU conditions of patients described in 1 to 9 any one The method of model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710310478.8A CN107436993B (en) | 2017-05-05 | 2017-05-05 | Method and server for establishing ICU patient condition evaluation model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710310478.8A CN107436993B (en) | 2017-05-05 | 2017-05-05 | Method and server for establishing ICU patient condition evaluation model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107436993A true CN107436993A (en) | 2017-12-05 |
| CN107436993B CN107436993B (en) | 2020-08-25 |
Family
ID=60458592
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710310478.8A Active CN107436993B (en) | 2017-05-05 | 2017-05-05 | Method and server for establishing ICU patient condition evaluation model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107436993B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109003670A (en) * | 2018-06-29 | 2018-12-14 | 武汉圣大东高科技有限公司 | Big data medical information processing method, system, terminal device and storage medium |
| CN109119155A (en) * | 2018-07-03 | 2019-01-01 | 厦门大学 | ICU mortality prediction assessment system based on deep learning |
| CN110010242A (en) * | 2019-04-03 | 2019-07-12 | 吉林大学第一医院 | A kind of intensive care monitoring system based on Internet of Things |
| CN110675956A (en) * | 2019-08-26 | 2020-01-10 | 南京医渡云医学技术有限公司 | Method and device for determining facial paralysis treatment scheme, readable medium and electronic equipment |
| CN111803804A (en) * | 2020-06-19 | 2020-10-23 | 山东省肿瘤防治研究院（山东省肿瘤医院） | Adaptive radiotherapy system, storage medium and apparatus |
| CN112349411A (en) * | 2020-12-03 | 2021-02-09 | 郑州大学第一附属医院 | ICU patient rescue risk prediction method and system based on big data |
| CN112365978A (en) * | 2020-11-10 | 2021-02-12 | 北京航空航天大学 | Method and device for establishing early risk assessment model of tachycardia event |
| CN112786166A (en) * | 2020-03-19 | 2021-05-11 | 中国医学科学院北京协和医院 | Data processing method, device, equipment and computer readable storage medium |
| CN113223703A (en) * | 2021-05-19 | 2021-08-06 | 中科彭州智慧产业创新中心有限公司 | Method and device for evaluating traditional Chinese medicine physiotherapy based on big data |
| CN113393934A (en) * | 2021-06-07 | 2021-09-14 | 义金(杭州)健康科技有限公司 | Health trend estimation method and prediction system based on vital sign big data |
| CN114694852A (en) * | 2022-04-13 | 2022-07-01 | 武汉科瓴智能科技有限公司 | Chronic disease analysis method and system |
| CN117789965A (en) * | 2024-02-23 | 2024-03-29 | 北京惠每云科技有限公司 | Venous thrombosis prevention simulation evaluation method and system based on digital twin |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103201743A (en) * | 2010-11-08 | 2013-07-10 | 皇家飞利浦电子股份有限公司 | Method of continuous prediction of patient severity of illness, mortality, and length of stay |
| CN103690240A (en) * | 2013-09-16 | 2014-04-02 | 上海华美络信息技术有限公司 | Medical system |
| CN105787232A (en) * | 2014-12-17 | 2016-07-20 | 中国移动通信集团公司 | Data processing method, device, health system platform and terminal |
| CN105956374A (en) * | 2016-04-21 | 2016-09-21 | 青岛大学 | Remote monitoring comprehensive medical system |
| CN106108921A (en) * | 2016-08-19 | 2016-11-16 | 北京中盛凯新科技发展有限公司 | A kind of psychosoma assessment and nursing and interfering system |
| CN106778042A (en) * | 2017-01-26 | 2017-05-31 | 中电科软件信息服务有限公司 | Cardio-cerebral vascular disease patient similarity analysis method and system |
-
2017
- 2017-05-05 CN CN201710310478.8A patent/CN107436993B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103201743A (en) * | 2010-11-08 | 2013-07-10 | 皇家飞利浦电子股份有限公司 | Method of continuous prediction of patient severity of illness, mortality, and length of stay |
| CN103690240A (en) * | 2013-09-16 | 2014-04-02 | 上海华美络信息技术有限公司 | Medical system |
| CN105787232A (en) * | 2014-12-17 | 2016-07-20 | 中国移动通信集团公司 | Data processing method, device, health system platform and terminal |
| CN105956374A (en) * | 2016-04-21 | 2016-09-21 | 青岛大学 | Remote monitoring comprehensive medical system |
| CN106108921A (en) * | 2016-08-19 | 2016-11-16 | 北京中盛凯新科技发展有限公司 | A kind of psychosoma assessment and nursing and interfering system |
| CN106778042A (en) * | 2017-01-26 | 2017-05-31 | 中电科软件信息服务有限公司 | Cardio-cerebral vascular disease patient similarity analysis method and system |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109003670A (en) * | 2018-06-29 | 2018-12-14 | 武汉圣大东高科技有限公司 | Big data medical information processing method, system, terminal device and storage medium |
| CN109119155A (en) * | 2018-07-03 | 2019-01-01 | 厦门大学 | ICU mortality prediction assessment system based on deep learning |
| CN109119155B (en) * | 2018-07-03 | 2022-01-28 | 厦门大学 | ICU death risk assessment system based on deep learning |
| CN110010242A (en) * | 2019-04-03 | 2019-07-12 | 吉林大学第一医院 | A kind of intensive care monitoring system based on Internet of Things |
| CN110675956A (en) * | 2019-08-26 | 2020-01-10 | 南京医渡云医学技术有限公司 | Method and device for determining facial paralysis treatment scheme, readable medium and electronic equipment |
| CN112786166A (en) * | 2020-03-19 | 2021-05-11 | 中国医学科学院北京协和医院 | Data processing method, device, equipment and computer readable storage medium |
| CN111803804A (en) * | 2020-06-19 | 2020-10-23 | 山东省肿瘤防治研究院（山东省肿瘤医院） | Adaptive radiotherapy system, storage medium and apparatus |
| CN112365978A (en) * | 2020-11-10 | 2021-02-12 | 北京航空航天大学 | Method and device for establishing early risk assessment model of tachycardia event |
| CN112349411B (en) * | 2020-12-03 | 2021-07-23 | 郑州大学第一附属医院 | ICU patient rescue risk prediction method and system based on big data |
| CN112349411A (en) * | 2020-12-03 | 2021-02-09 | 郑州大学第一附属医院 | ICU patient rescue risk prediction method and system based on big data |
| CN113223703A (en) * | 2021-05-19 | 2021-08-06 | 中科彭州智慧产业创新中心有限公司 | Method and device for evaluating traditional Chinese medicine physiotherapy based on big data |
| CN113393934A (en) * | 2021-06-07 | 2021-09-14 | 义金(杭州)健康科技有限公司 | Health trend estimation method and prediction system based on vital sign big data |
| CN113393934B (en) * | 2021-06-07 | 2022-07-12 | 义金(杭州)健康科技有限公司 | Health trend estimation method and prediction system based on vital sign big data |
| CN114694852A (en) * | 2022-04-13 | 2022-07-01 | 武汉科瓴智能科技有限公司 | Chronic disease analysis method and system |
| CN117789965A (en) * | 2024-02-23 | 2024-03-29 | 北京惠每云科技有限公司 | Venous thrombosis prevention simulation evaluation method and system based on digital twin |
| CN117789965B (en) * | 2024-02-23 | 2024-06-07 | 北京惠每云科技有限公司 | Venous thrombosis prevention simulation evaluation method and system based on digital twin |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107436993B (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107436993A (en) | Establish the method and server of ICU conditions of patients assessment models | |
| US20220254486A1 (en) | System and method for a patient dashboard | |
| CN110459328A (en) | A kind of Clinical Decision Support Systems for assessing sudden cardiac arrest | |
| CN111971756A (en) | Health assistance system, information providing form output device, method, and program | |
| JP2023123711A (en) | Prognostic score based on health information | |
| Grachev et al. | Structural abnormalities of frontal neocortex in obsessive-compulsive disorder | |
| CN111048167A (en) | Hierarchical case structuring method and system | |
| Castiñeira et al. | Adding continuous vital sign information to static clinical data improves the prediction of length of stay after intubation: a data-driven machine learning approach | |
| CN107391901A (en) | Establish the method and server of public ward conditions of patients assessment models | |
| CN112967803A (en) | Early mortality prediction method and system for emergency patients based on integrated model | |
| CN116895372A (en) | Intelligent first-aid grading system based on large-scale language model and meta-learning | |
| CN117198543A (en) | Wound resuscitation unit complication prediction model training and prediction method and electronic equipment | |
| CN115862850A (en) | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data | |
| US20210407676A1 (en) | Patient ventilator asynchrony detection | |
| Faisal et al. | A machine learning approach for analyzing and predicting suicidal thoughts and behaviors | |
| JP7044113B2 (en) | Presentation method, presentation system, and program | |
| Orfanoudaki et al. | Algorithm, Human, or the Centaur: How to Enhance Clinical Care? | |
| Hovorushchenko et al. | Supporting the decision-making about the possibility of donation and transplantation based on civil law grounds | |
| US20210282649A1 (en) | Systems and methods for predicting high frequency and low frequency patient parameters | |
| US20170140115A1 (en) | Intelligent stroke risk prediction and monitoring system | |
| US20190287675A1 (en) | Systems and methods for determining healthcare quality measures by evalutating subject healthcare data in real-time | |
| Loperto et al. | Use of regression models to predict glomerular filtration rate in kidney transplanted patients | |
| Khalaf et al. | Applied difference techniques of machine learning algorithm and web-based management system for sickle cell disease | |
| Siu et al. | An Intelligent clinical decision support system for assessing the needs of a long-term care plan | |
| CN110910980A (en) | Sepsis early warning device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |