CN109411082A - A kind of Evaluation of Medical Quality and medical recommended method - Google Patents
A kind of Evaluation of Medical Quality and medical recommended method Download PDFInfo
- Publication number
- CN109411082A CN109411082A CN201811322329.4A CN201811322329A CN109411082A CN 109411082 A CN109411082 A CN 109411082A CN 201811322329 A CN201811322329 A CN 201811322329A CN 109411082 A CN109411082 A CN 109411082A
- Authority
- CN
- China
- Prior art keywords
- medical
- recommended
- evaluation
- case
- patient
- 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
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a kind of Evaluation of Medical Quality and medical recommended methods, which comprises step S1 obtains the data for evaluating quality of medical care；Step S2 evaluates the data prediction of quality of medical care；Step S3, case grouping；Step S4 segments group cluster by case and draws dendrogram；Step S5 calculates the Evaluation of Medical Quality point of case subdivision group according to dendrogram；Step S6 calculates the Evaluation of Medical Quality point of case rough segmentation group；Step S7, the final Evaluation of Medical Quality point of Calculation Estimation object；Step S8 obtains for training and tests the data set of medical recommended models；Step S9, building obtain the neural network model of patient characteristic vector；Step S10, building obtain the neural network model of recommended feature vector；Step S11 constructs medical recommendation neural network model；Step S12, model training；Step S13 generates medical recommendation list, recommends patient.
Description
Technical field
The present invention relates to medical big data fields, more particularly to a kind of Evaluation of Medical Quality and medical recommended method.
Background technique
Medical safety and quality of medical care are all the hot issue paid much attention to both at home and abroad for a long time.Doctor of the country to hospital
Treatment level and criteria of quality evaluation are " three-level general hospital review standards of designing (version in 2011) ", and " three-level general hospital medical treatment
Quality management and control standard (2011 editions) ", evaluating standard includes: Death class index, returns to class index, nosocomial infection class
Index, postoperative complication class index, patient safety class index, medical institutions' Rational Use of Drug Index and hospital's operational management class refer to
Mark.
Documents and materials show that common Evaluation of Medical Quality method can be summarized as Hospital Accreditation method, comprehensive evaluation and review technique, disease
Kind quality evaluation and review technique and customer satisfaction evaluation and review technique etc..The common feature of these methods is that evaluation procedure requires manual intervention and sentences
It is disconnected, depend on the experience of expert unduly；Meanwhile analysis and evaluation system is to be based on random sample rather than all data.Therefore, medical
The defect in quality evaluation field first is that the evaluation of medical institutions, doctor etc. there are lacks objectivity, scientific insufficient ask
Topic.Currently, generally there are no carry out objective comparison, nothing using medical method of the technological means such as big data analysis to doctor for hospital
Method judges its superiority and inferiority, and the treatment method of the doctor caused cannot be promoted, and hospital quality management is caused to there is careless omission, reduces
The efficiency of management and quality.Meanwhile in treatment process, patient selects the mode of department or doctor generally by kith and kin, hangs
Number member or overt propaganda information, these channels lack the objective description and ratio to department's diagnosis and treatment characteristic and physician specialty speciality
Compared with so as to cause hospital treatment services cannot be provided timely, high-qualityly for patient.
Summary of the invention
The technical problems to be solved by the present invention are: objective for current Evaluation of Medical Quality and medical recommended method shortage
Property, scientific problem, be based on hospital journals big data, innovatively design a kind of Evaluation of Medical Quality and medical recommended method,
So as to Comprehensive promote medical quality in hospital management level.
To solve the above-mentioned problems, the invention discloses a kind of Evaluation of Medical Quality and medical recommended method, technical sides
Case the following steps are included:
Step S1, obtains the data for evaluating quality of medical care, including patient ID, Gender, patient age, department of being admitted to hospital,
Attending physician, situation of being admitted to hospital, admission diagnosis code, therapeutic effect, treatment time and medical expense etc.；
Step S2 evaluates the data prediction of quality of medical care；
Step S3, case grouping carry out rough segmentation group to case according to admission diagnosis code first, then comprehensively consider the property of patient
Not, age and situation of being admitted to hospital are finely divided group；
Step S4, for each case subdivision group, using treatment time, medical expense, therapeutic effect as evaluation index；First
Evaluation index is standardized；Then the classification number k value of cluster is determined using error sum of squares SSE index；It is finally right
Evaluation index is clustered, and case is sorted out, and draws dendrogram；
Step S5 calculates each evaluation object in the Evaluation of Medical Quality point of case subdivision group according to dendrogram；
Step S6 divides weighted sum to Evaluation of Medical Quality of each evaluation object in each case subdivision group, obtains each
Case rough segmentation group is the Evaluation of Medical Quality point being grouped according to admission diagnosis code；
Step S7, by Evaluation of Medical Quality of each evaluation object in each case rough segmentation group point according to sorting from high to low,
The case type that the case of n evaluation point is good at as the evaluation object before taking, to the doctor for the case type that evaluation object is good at
It treats quality evaluation and divides weighted sum, obtain the final Evaluation of Medical Quality point of each evaluation object；
Step S8 obtains for training and tests the data set of medical recommended models；Data set is made of three files, is respectively
Patient data files, recommended data file and score data file；Patient data files include patient's id field, sexual
Malapropism section, patient age field and illness description field；Recommended data file includes recommended id field and recommendation pair
As the case type field being good at；Score data file include patient's id field, recommended id field and scoring field, wherein
The value for the field that scores is training and the target value for testing medical recommended models；
Step S9, building obtain the neural network model of patient characteristic vector；
Step S10, building obtain the neural network model of recommended feature vector；
Step S11 constructs medical recommendation neural network model；
Data set is divided into training set and test set by step S12, is lost using mean square error MSE index optimization, determines that training changes
Generation number and learning rate carry out model training, obtain patient characteristic matrix and recommended eigenmatrix；
Step S13 calculates the score value of object to be recommended using recommended eigenmatrix described in the patient characteristic vector sum,
It takes score value highest first n to generate medical recommendation list, recommends patient.
A kind of Evaluation of Medical Quality and medical recommended method, the step S5 further include:
Step S51, the evaluation object, including doctor, department；
Step S52 counts case number of each evaluation object in each cluster classification of case subdivision group, by what is selected
The case number in classification is clustered divided by total case number, obtains each evaluation object in the quality of medical care of case subdivision group
Evaluation point.
A kind of Evaluation of Medical Quality and medical recommended method, the step S8 further include:
Step S81, the recommended is consistent with evaluation object described in the step S51, including doctor, department；
Step S82, the illness description field carry out word segmentation processing to the text description in the illness description field first,
Then word is established to the mapping dictionary of number, converts thereof into numerical listing, and the length of all numerical listings is unified；
Step S83, the value for the case type field that the recommended is good at can be obtained, it may be assumed that will be every by the step S7
Evaluation of Medical Quality of a evaluation object in each case rough segmentation group point according to sorting from high to low, and n evaluation divides before taking
The case type that case is good at as the recommended；Establish the class malapropism in the case type field that the recommended is good at
Symbol string converts thereof into numerical listing to the mapping dictionary of number, and the length of all numerical listings is unified；
Step S84, the value of the scoring field can be obtained, it may be assumed that thin according to case by the step S4 and the step S5
The result of grouping and clustering and selected cluster classification set 1 for the selected corresponding score value of cluster classification, remaining is 0.
A kind of Evaluation of Medical Quality and medical recommended method, the step S9 further include:
Step S91, the first layer of the neural network model are embeding layer, by patient's id field, Gender field, are suffered from
Person's age field is passed to embeding layer as the index of embeded matrix, obtains the patient ID feature, Gender feature and patient
Age characteristics；
Step S92 is handled for the illness description field using text convolutional neural networks；First from embeded matrix
The insertion vector of the corresponding each word of the illness description field is obtained, then carries out convolution fortune using various sizes of convolution kernel
Calculation, maximum pond and regularization operation, obtain the feature of the illness description field；
It is special to index out the patient ID feature, Gender feature and patient age from the neural network embeding layer by step S93
It levies and after the feature that text convolutional neural networks obtain the illness description field, each feature incoming one is connected entirely
Layer is connect, outputs it and is passed to a full articulamentum again, to obtain the patient characteristic vector.
A kind of Evaluation of Medical Quality and medical recommended method, the step S10 further include:
The first layer of step S101, the neural network model are embeding layer, by the recommended id field and recommended
The case type field being good at is passed to embeding layer as the index of embeded matrix, obtains the recommended ID feature and recommendation pair
As the case type feature being good at；
Step S102 is further processed the case type feature that the recommended is good at, i.e., arrogates to oneself the recommended
Multiple insertion vectors of long case type do adduction operation；
Step S103, from the neural network embeding layer indexes out the recommended ID feature and the recommended is good at
Case type feature and after being further processed to the case type feature that recommended is good at, each feature is one incoming
Full articulamentum outputs it and is passed to a full articulamentum again, obtains the recommended feature vector.
A kind of Evaluation of Medical Quality and medical recommended method, the step S11 further include:
Recommended feature vector described in the patient characteristic vector sum is done vector multiplication by step S111, by calculated result with
True score value returns.
Compared with prior art, the invention has the following advantages that
(1) present invention can carry out objective ratio to the quality of medical care of different doctors, department using hospital's this case of bulk sample big data
Compared with, quantify the level professional technology and medical service quality of doctor, department, it being capable of General Promotion medical quality in hospital management water
Flat, evaluation result can be used as the decision-making foundation of patient assessment's selection.
(2) present invention can analyze doctor, the case type that department is good at, precise positioning doctor, the feature of department and thin
Weak link improves medical level so that hospital can take specific aim measure.
(3) in daily treatment process, patient select department, doctor mode generally by kith and kin, Registrar or
Overt propaganda information, these channels lack the objective description to department's diagnosis and treatment characteristic and physician specialty speciality and compare, thus
Cause hospital that cannot provide treatment services timely, high-qualityly for patient.In addition to this, some documents are proposed through acquisition from mutual
The evaluation information of networking recommends doctor, department, due to the subjectivity of evaluation, one-sidedness and the sparsity for evaluating data, meeting
Seriously affect the effect of recommendation.The present invention uses big data analysis technology, objectively evaluates to the quality of medical care of doctor, department
Marking, can obtain in time doctor, the case type that department is good at, objectively respond doctor, the case type that department is good at it is dynamic
State variation realizes precisely medical recommend by deep neural network.
(4) document shows that medical recommender system generally uses collaborative filtering, the history that this method passes through acquisition patient
The information such as record, personal preference are calculated the similarity with other patients, are recommended doctor, department using the evaluation of similar patients.
In general, there is cold start-up in collaborative filtering.The present invention obtains the neural network mould of patient characteristic vector by building
Type and building obtain the neural network model of recommended feature vector, recommend neural network model to realize and go to a doctor, and use
Recommended eigenmatrix described in the patient characteristic vector sum calculates the score value of object to be recommended, realizes precisely medical push away
It recommends.
Detailed description of the invention
Fig. 1 is the flow chart of a kind of Evaluation of Medical Quality and medical recommended method of the invention.
Specific embodiment
As shown in Fig. 1, the method for the present invention follows the steps below:
The present invention is described in detail with reference to the accompanying drawing.
Step S1 obtains the data for evaluating quality of medical care, including patient ID, Gender, patient age, section of being admitted to hospital
Room, attending physician, situation of being admitted to hospital, admission diagnosis code, therapeutic effect, treatment time and medical expense etc.；
Since the 1990s, hospital has been gradually completing clinical treatment informationization, has built up HIS(with the stream of people, logistics, wealth
Flow management is the hospital information system of core), EMRs(is for the purpose of the information collection of patient medical record, storage and centralized management
Electronic medical record system), LIS(checking information system), PACS(medical image information system), NIS(nursing information system) etc..Cause
This, for evaluating the patient ID of quality of medical care, Gender, patient age, department of being admitted to hospital, attending physician, situation of being admitted to hospital, being admitted to hospital
The data such as diagnosis code, therapeutic effect, treatment time and medical expense can be extracted from above-mentioned medical information system.
Step S2 evaluates the data prediction of quality of medical care；
Since medical information system is gradually built by stages, in addition the reasons such as case database upgrading, there are data
The problems such as Statistical Criteria is inconsistent, the overlapping of data format disunity, data information, noise data and data are omitted, it is necessary to logarithm
According to the operation such as being cleaned, being extracted, being converted, being arranged and being filled, to realize the standardization and standardization of data format, such as: NaN
Processing, the processing of 0 value of medical expense and type of name conversion of value etc..
Step S3, case grouping carry out rough segmentation group to case according to admission diagnosis code first, then comprehensively consider patient's
Gender, age and situation of being admitted to hospital are finely divided group；
Rough segmentation group is carried out to case according to admission diagnosis code, to eliminate the influence of disease type difference bring.Then comprehensively consider
Gender, age and the situation of being admitted to hospital of patient is finely divided group, is brought with eliminating the differences such as case personal feature and medical history
Influence so that quantizing process can, resource consumption comparable case close to clinical process evaluate.
Step S4, for each case subdivision group, using treatment time, medical expense, therapeutic effect as evaluation index；
Evaluation index is standardized first；Then the classification number k value of cluster is determined using error sum of squares SSE index；Most
Evaluation index is clustered afterwards, case is sorted out, and draws dendrogram；
Using treatment time, medical expense, therapeutic effect as quality of medical care quantizating index, sufferer is not only reflected jointly because for the treatment of
And " cost " is paid, and largely reflect the height for the treatment of level, so as to fully demonstrate quality of medical care
Superiority and inferiority；
Since the magnitude differences between each evaluation index are larger, cluster effect is influenced in order to avoid the value of single feature is excessive
Fruit needs to be standardized before being clustered；
After calculating error sum of squares SSE and determining the classification number k value clustered, is clustered, case is sorted out.Meanwhile for poly-
Class result is by the different classes of line chart for drawing feature.
Step S5 calculates each evaluation object in the Evaluation of Medical Quality point of case subdivision group according to dendrogram；
The evaluation object, including doctor, department；
Case number of each evaluation object in each cluster classification of case subdivision group is counted, by selected cluster classification
In case number divided by total case number, obtain each evaluation object in the Evaluation of Medical Quality point of case subdivision group；
For the specific embodiment that the present invention will be described in detail, be exemplified below: hypothesis evaluation object 1 is in treatment case subdivision
There are 91,1031 and 9 cases respectively in 3 cluster classifications 0,1,2 of group G, by the characteristic folding lines for analyzing different cluster classifications
Figure wherein cluster classification 1 is the cluster classification that therapeutic effect is good, treatment time is short, medical expense is low, therefore selects cluster classification
1 for calculating Evaluation of Medical Quality point, then evaluation object 1 is divided into 1031/ in the Evaluation of Medical Quality for the treatment of case subdivision group G
(91+1031+9)=91.2%。
Step S6 divides weighted sum to Evaluation of Medical Quality of each evaluation object in each case subdivision group, obtains
Each case rough segmentation group is the Evaluation of Medical Quality point being grouped according to admission diagnosis code；
For the specific embodiment that the present invention will be described in detail, be exemplified below: hypothesis evaluation object 1 is segmented in i case
Evaluation of Medical Quality point in group is respectively s1, s2 ..., si, then evaluation object 1 is commented in the quality of medical care of some case rough segmentation group
Valence is divided into r=w1*s1+w2*s2+ ...+wi*si, wherein w1, w2 ..., wi are weight, are met: 0 < w1 < 1,0 < w2 < 1 ..., 0 < wi
< 1, and w1+w2+ ...+wi=1.
Step S7, by Evaluation of Medical Quality of each evaluation object in each case rough segmentation group point according to arranging from high to low
Sequence, the case type be good at as the evaluation object of case that n evaluation divides before taking, the case type that evaluation object is good at
Evaluation of Medical Quality divides weighted sum, obtains the final Evaluation of Medical Quality point of each evaluation object；
For the specific embodiment that the present invention will be described in detail, be exemplified below: hypothesis evaluation object 1 is in j case rough segmentation
Evaluation of Medical Quality point in group by r1, r2 ..., rj is ordered as from high to low, the n disease being good at as evaluation object before taking
Example type, then the final Evaluation of Medical Quality of evaluation object 1 is divided into q=w1*r1+w2*r2+ ...+wn*rn, wherein w1, w2 ...,
Wn is weight, is met: 0<w1<1,0<w2<1 ..., 0<wn<1, and w1>w2>...+wn=1>wn, w1+w2+ ....
Step S8 obtains for training and tests the data set of medical recommended models；Data set is made of three files, point
It is not patient data files, recommended data file and score data file；Patient data files include patient's id field, suffer from
Person's gender field, patient age field and illness description field；Recommended data file includes recommended id field and pushes away
Recommend the case type field that object is good at；Score data file include patient's id field, recommended id field and scoring field,
The value for the field that wherein scores is training and the target value for testing medical recommended models；
The recommended, it is consistent with the evaluation object, including doctor, department；
The illness description field carries out word segmentation processing to the text description in the illness description field first, then establishes
Word converts thereof into numerical listing to the mapping dictionary of number, and the length of all numerical listings is unified；
The value for the case type field that the recommended is good at can be obtained by the step S7, it may be assumed that by each evaluation pair
As the case conduct that the Evaluation of Medical Quality in each case rough segmentation group point according to sorting from high to low, and n evaluation divides before taking
The case type that the recommended is good at；Classification character string in the case type field that the recommended is good at is established to number
The mapping dictionary of word converts thereof into numerical listing, and the length of all numerical listings is unified；
The value of the scoring field, can be obtained by the step S4 and the step S5, it may be assumed that segment group cluster according to case
Result and selected cluster classification, set 1 for the selected corresponding score value of cluster classification, remaining is 0, wherein 1 indicates
Scoring of the patient that therapeutic effect is good in case subdivision group, treatment time is short, medical expense is low to recommended.
Step S9, building obtain the neural network model of patient characteristic vector；
The first layer of the neural network model is embeding layer, by patient's id field, Gender field, patient age word
The index of Duan Dangzuo embeded matrix is passed to embeding layer, obtains the patient ID feature, Gender feature and patient age feature；
It is handled for the illness description field using text convolutional neural networks；It is obtained from embeded matrix first described
Then the insertion vector of the corresponding each word of illness description field carries out convolution algorithm, maximum using various sizes of convolution kernel
Pondization and regularization operate, and obtain the feature of the illness description field；
From the neural network embeding layer index out the patient ID feature, Gender feature and patient age feature and from
After text convolutional neural networks obtain the feature of the illness description field, each feature is passed to a full articulamentum, it will
It is exported is passed to a full articulamentum again, to obtain the patient characteristic vector.
Step S10, building obtain the neural network model of recommended feature vector；
The first layer of the neural network model is embeding layer, the case that the recommended id field and recommended are good at
Type field is passed to embeding layer as the index of embeded matrix, obtains the disease that the recommended ID feature and recommended are good at
Example type feature；
The case type feature that the recommended is good at is further processed, i.e., the case class recommended being good at
Multiple insertion vectors of type do adduction operation；
The case type that the recommended ID feature and the recommended are good at is indexed out from the neural network embeding layer
Feature and after being further processed to the case type feature that recommended is good at, by the incoming one full connection of each feature
Layer outputs it and is passed to a full articulamentum again, obtains the recommended feature vector.
Step S11 constructs medical recommendation neural network model；
Recommended feature vector described in the patient characteristic vector sum is done into vector multiplication, by calculated result and true scoring
Value returns.
Data set is divided into training set and test set by step S12, is lost using mean square error MSE index optimization, determines instruction
Practice the number of iterations and learning rate, carries out model training, obtain patient characteristic matrix and recommended eigenmatrix；
For training pattern, it is necessary first to define cost or loss function carrys out assessment models, be referred to here using mean square error MSE
Mark；Then it needs to be determined that training the number of iterations and learning rate；Trained process is exactly continuous adjustment model parameter, is minimized equal
Square error MSE index.
Step S13 calculates commenting for object to be recommended using recommended eigenmatrix described in the patient characteristic vector sum
Score value takes score value highest first n to generate medical recommendation list, recommends patient.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical scheme and beneficial effects
It describes in detail bright.Those skilled in the art is not under conditions of departing from the spirit and scope of the present invention that claims determine, also
Various modifications can be carried out to the above content.Therefore, the scope of the present invention is not limited in above explanation, but by
The range of claims determines.
Claims (6)
1. a kind of Evaluation of Medical Quality and medical recommended method characterized by comprising
Step S1, obtains the data for evaluating quality of medical care, including patient ID, Gender, patient age, department of being admitted to hospital,
Attending physician, situation of being admitted to hospital, admission diagnosis code, therapeutic effect, treatment time and medical expense etc.；
Step S2 evaluates the data prediction of quality of medical care；
Step S3, case grouping carry out rough segmentation group to case according to admission diagnosis code first, then comprehensively consider the property of patient
Not, age and situation of being admitted to hospital are finely divided group；
Step S4, for each case subdivision group, using treatment time, medical expense, therapeutic effect as evaluation index；First
Evaluation index is standardized；Then the classification number k value of cluster is determined using error sum of squares SSE index；It is finally right
Evaluation index is clustered, and case is sorted out, and draws dendrogram；
Step S5 calculates each evaluation object in the Evaluation of Medical Quality point of case subdivision group according to dendrogram；
Step S6 divides weighted sum to Evaluation of Medical Quality of each evaluation object in each case subdivision group, obtains each
Case rough segmentation group is the Evaluation of Medical Quality point being grouped according to admission diagnosis code；
Step S7, by Evaluation of Medical Quality of each evaluation object in each case rough segmentation group point according to sorting from high to low,
The case type that the case of n evaluation point is good at as the evaluation object before taking, to the doctor for the case type that evaluation object is good at
It treats quality evaluation and divides weighted sum, obtain the final Evaluation of Medical Quality point of each evaluation object；
Step S8 obtains for training and tests the data set of medical recommended models；Data set is made of three files, is respectively
Patient data files, recommended data file and score data file；Patient data files include patient's id field, sexual
Malapropism section, patient age field and illness description field；Recommended data file includes recommended id field and recommendation pair
As the case type field being good at；Score data file include patient's id field, recommended id field and scoring field, wherein
The value for the field that scores is training and the target value for testing medical recommended models；
Step S9, building obtain the neural network model of patient characteristic vector；
Step S10, building obtain the neural network model of recommended feature vector；
Step S11 constructs medical recommendation neural network model；
Data set is divided into training set and test set by step S12, is lost using mean square error MSE index optimization, determines that training changes
Generation number and learning rate carry out model training, obtain patient characteristic matrix and recommended eigenmatrix；
Step S13 calculates the score value of object to be recommended using recommended eigenmatrix described in the patient characteristic vector sum,
It takes score value highest first n to generate medical recommendation list, recommends patient.
2. Evaluation of Medical Quality according to claim 1 and medical recommended method, which is characterized in that the step S5 is also wrapped
It includes:
Step S51, the evaluation object, including doctor, department；
Step S52 counts case number of each evaluation object in each cluster classification of case subdivision group, by what is selected
The case number in classification is clustered divided by total case number, obtains each evaluation object in the quality of medical care of case subdivision group
Evaluation point.
3. Evaluation of Medical Quality according to claim 1 and medical recommended method, which is characterized in that the step S8 is also wrapped
It includes:
Step S81, the recommended is consistent with evaluation object described in the step S51, including doctor, department；
Step S82, the illness description field carry out word segmentation processing to the text description in the illness description field first,
Then word is established to the mapping dictionary of number, converts thereof into numerical listing, and the length of all numerical listings is unified；
Step S83, the value for the case type field that the recommended is good at can be obtained, it may be assumed that will be every by the step S7
Evaluation of Medical Quality of a evaluation object in each case rough segmentation group point according to sorting from high to low, and n evaluation divides before taking
The case type that case is good at as the recommended；Establish the class malapropism in the case type field that the recommended is good at
Symbol string converts thereof into numerical listing to the mapping dictionary of number, and the length of all numerical listings is unified；
Step S84, the value of the scoring field can be obtained, it may be assumed that thin according to case by the step S4 and the step S5
The result of grouping and clustering and selected cluster classification set 1 for the selected corresponding score value of cluster classification, remaining is 0.
4. Evaluation of Medical Quality according to claim 1 and medical recommended method, which is characterized in that the step S9 is also wrapped
It includes:
Step S91, the first layer of the neural network model are embeding layer, by patient's id field, Gender field, are suffered from
Person's age field is passed to embeding layer as the index of embeded matrix, obtains the patient ID feature, Gender feature and patient
Age characteristics；
Step S92 is handled for the illness description field using text convolutional neural networks；First from embeded matrix
The insertion vector of the corresponding each word of the illness description field is obtained, then carries out convolution fortune using various sizes of convolution kernel
Calculation, maximum pond and regularization operation, obtain the feature of the illness description field；
It is special to index out the patient ID feature, Gender feature and patient age from the neural network embeding layer by step S93
It levies and after the feature that text convolutional neural networks obtain the illness description field, each feature incoming one is connected entirely
Layer is connect, outputs it and is passed to a full articulamentum again, to obtain the patient characteristic vector.
5. Evaluation of Medical Quality according to claim 1 and medical recommended method, which is characterized in that the step S10 is also
Include:
The first layer of step S101, the neural network model are embeding layer, by the recommended id field and recommended
The case type field being good at is passed to embeding layer as the index of embeded matrix, obtains the recommended ID feature and recommendation pair
As the case type feature being good at；
Step S102 is further processed the case type feature that the recommended is good at, i.e., arrogates to oneself the recommended
Multiple insertion vectors of long case type do adduction operation；
Step S103, from the neural network embeding layer indexes out the recommended ID feature and the recommended is good at
Case type feature and after being further processed to the case type feature that recommended is good at, each feature is one incoming
Full articulamentum outputs it and is passed to a full articulamentum again, obtains the recommended feature vector.
6. Evaluation of Medical Quality according to claim 1 and medical recommended method, which is characterized in that the step S11 is also
Include:
Recommended feature vector described in the patient characteristic vector sum is done vector multiplication by step S111, by calculated result with
True score value returns.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811322329.4A CN109411082B (en) | 2018-11-08 | 2018-11-08 | Medical quality evaluation and treatment recommendation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811322329.4A CN109411082B (en) | 2018-11-08 | 2018-11-08 | Medical quality evaluation and treatment recommendation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109411082A true CN109411082A (en) | 2019-03-01 |
| CN109411082B CN109411082B (en) | 2022-01-04 |
Family
ID=65472262
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811322329.4A Active CN109411082B (en) | 2018-11-08 | 2018-11-08 | Medical quality evaluation and treatment recommendation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109411082B (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110310745A (en) * | 2019-05-21 | 2019-10-08 | 上海交通大学医学院附属瑞金医院 | The therapeutic scheme recommender system that medical guide and data-driven combine |
| CN110322959A (en) * | 2019-05-24 | 2019-10-11 | 山东大学 | A kind of Knowledge based engineering depth medical care problem method for routing and system |
| CN110516040A (en) * | 2019-08-14 | 2019-11-29 | 出门问问(武汉)信息科技有限公司 | Semantic Similarity comparative approach, equipment and computer storage medium between text |
| CN110752017A (en) * | 2019-09-04 | 2020-02-04 | 重庆特斯联智慧科技股份有限公司 | Community doctor scheduling method and system based on deep learning |
| CN110888887A (en) * | 2019-12-09 | 2020-03-17 | 湖南新云医疗装备工业有限公司 | Intelligent display method and system for medical care information |
| CN111048213A (en) * | 2019-12-06 | 2020-04-21 | 张彩东 | Emergency call quality assessment management system |
| CN111091898A (en) * | 2019-11-14 | 2020-05-01 | 泰康保险集团股份有限公司 | Medical institution evaluation system, method, device, storage medium and electronic equipment |
| CN111161814A (en) * | 2019-12-18 | 2020-05-15 | 浙江大学 | DRGs automatic grouping method based on convolutional neural network |
| CN111507414A (en) * | 2020-04-20 | 2020-08-07 | 安徽中科首脑智能医疗研究院有限公司 | Deep learning skin disease picture comparison and classification method, storage medium and robot |
| CN112035741A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Reservation method, device, equipment and storage medium based on user physical examination data |
| CN113539460A (en) * | 2021-07-29 | 2021-10-22 | 深圳万海思数字医疗有限公司 | Intelligent diagnosis guiding method and device for remote medical platform |
| CN113555094A (en) * | 2020-04-24 | 2021-10-26 | 安徽科大讯飞医疗信息技术有限公司 | Hospital evaluation method, device, equipment and medium |
| CN114334176A (en) * | 2020-09-30 | 2022-04-12 | 西门子医疗有限公司 | Computer-implemented method, device and medical system |
| CN114722977A (en) * | 2022-06-10 | 2022-07-08 | 四川大学 | Medical object classification method and device, electronic equipment and storage medium |
| CN116386856A (en) * | 2023-06-05 | 2023-07-04 | 之江实验室 | Multi-label disease auxiliary diagnosis system based on doctor decision mode identification |
| CN116434954A (en) * | 2023-03-08 | 2023-07-14 | 绍兴珂西生物科技有限公司 | Evaluation method of clinical treatment effect |
| EP4038617A4 (en) * | 2019-10-02 | 2023-11-29 | Endpoint Health Inc. | Directing medical diagnosis and intervention recommendations |
| CN117558461A (en) * | 2024-01-12 | 2024-02-13 | 四川互慧软件有限公司 | Similar snake bite medical scheme selection method and device in different regions and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202306554U (en) * | 2011-07-22 | 2012-07-04 | 大连亿创天地科技发展有限公司 | Doctor evaluation grading system based on network |
| CN106202891A (en) * | 2016-06-30 | 2016-12-07 | 电子科技大学 | A kind of big data digging method towards Evaluation of Medical Quality |
| CN106934018A (en) * | 2017-03-11 | 2017-07-07 | 广东省中医院 | A kind of doctor's commending system based on collaborative filtering |
| CN107436933A (en) * | 2017-07-20 | 2017-12-05 | 广州慧扬健康科技有限公司 | The hierarchical clustering system arranged for case history archive |
| CN107463770A (en) * | 2017-07-11 | 2017-12-12 | 武汉金豆医疗数据科技有限公司 | A kind of evaluation method and system based on medical diagnosis on disease associated packets |
| US20170372028A1 (en) * | 2016-06-22 | 2017-12-28 | Xerox Corporation | System and method for scoring the performance of healthcare organizations |
-
2018
- 2018-11-08 CN CN201811322329.4A patent/CN109411082B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202306554U (en) * | 2011-07-22 | 2012-07-04 | 大连亿创天地科技发展有限公司 | Doctor evaluation grading system based on network |
| US20170372028A1 (en) * | 2016-06-22 | 2017-12-28 | Xerox Corporation | System and method for scoring the performance of healthcare organizations |
| CN106202891A (en) * | 2016-06-30 | 2016-12-07 | 电子科技大学 | A kind of big data digging method towards Evaluation of Medical Quality |
| CN106934018A (en) * | 2017-03-11 | 2017-07-07 | 广东省中医院 | A kind of doctor's commending system based on collaborative filtering |
| CN107463770A (en) * | 2017-07-11 | 2017-12-12 | 武汉金豆医疗数据科技有限公司 | A kind of evaluation method and system based on medical diagnosis on disease associated packets |
| CN107436933A (en) * | 2017-07-20 | 2017-12-05 | 广州慧扬健康科技有限公司 | The hierarchical clustering system arranged for case history archive |
Non-Patent Citations (2)
| Title |
|---|
| 李娟: "面向临床路径的常用数据挖掘方法概述", 《中国卫生产业》 * |
| 莫春梅 等: "医疗质量及其评价", 《中国医院统计》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110310745A (en) * | 2019-05-21 | 2019-10-08 | 上海交通大学医学院附属瑞金医院 | The therapeutic scheme recommender system that medical guide and data-driven combine |
| CN110310745B (en) * | 2019-05-21 | 2021-12-03 | 上海交通大学医学院附属瑞金医院 | Treatment plan recommendation system combining medical guide and data drive |
| CN110322959A (en) * | 2019-05-24 | 2019-10-11 | 山东大学 | A kind of Knowledge based engineering depth medical care problem method for routing and system |
| CN110322959B (en) * | 2019-05-24 | 2021-09-28 | 山东大学 | Deep medical problem routing method and system based on knowledge |
| CN110516040A (en) * | 2019-08-14 | 2019-11-29 | 出门问问(武汉)信息科技有限公司 | Semantic Similarity comparative approach, equipment and computer storage medium between text |
| CN110752017B (en) * | 2019-09-04 | 2020-12-18 | 重庆特斯联智慧科技股份有限公司 | Community doctor scheduling method and system based on deep learning |
| CN110752017A (en) * | 2019-09-04 | 2020-02-04 | 重庆特斯联智慧科技股份有限公司 | Community doctor scheduling method and system based on deep learning |
| EP4038617A4 (en) * | 2019-10-02 | 2023-11-29 | Endpoint Health Inc. | Directing medical diagnosis and intervention recommendations |
| CN111091898B (en) * | 2019-11-14 | 2023-08-22 | 泰康保险集团股份有限公司 | Medical institution evaluation system, method, device, storage medium and electronic apparatus |
| CN111091898A (en) * | 2019-11-14 | 2020-05-01 | 泰康保险集团股份有限公司 | Medical institution evaluation system, method, device, storage medium and electronic equipment |
| CN111048213A (en) * | 2019-12-06 | 2020-04-21 | 张彩东 | Emergency call quality assessment management system |
| CN110888887A (en) * | 2019-12-09 | 2020-03-17 | 湖南新云医疗装备工业有限公司 | Intelligent display method and system for medical care information |
| CN111161814A (en) * | 2019-12-18 | 2020-05-15 | 浙江大学 | DRGs automatic grouping method based on convolutional neural network |
| CN111507414A (en) * | 2020-04-20 | 2020-08-07 | 安徽中科首脑智能医疗研究院有限公司 | Deep learning skin disease picture comparison and classification method, storage medium and robot |
| CN113555094A (en) * | 2020-04-24 | 2021-10-26 | 安徽科大讯飞医疗信息技术有限公司 | Hospital evaluation method, device, equipment and medium |
| CN113555094B (en) * | 2020-04-24 | 2024-04-09 | 讯飞医疗科技股份有限公司 | Hospital assessment method, device, equipment and medium |
| CN112035741B (en) * | 2020-08-28 | 2022-08-30 | 康键信息技术（深圳）有限公司 | Reservation method, device, equipment and storage medium based on user physical examination data |
| CN112035741A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Reservation method, device, equipment and storage medium based on user physical examination data |
| CN114334176A (en) * | 2020-09-30 | 2022-04-12 | 西门子医疗有限公司 | Computer-implemented method, device and medical system |
| CN113539460A (en) * | 2021-07-29 | 2021-10-22 | 深圳万海思数字医疗有限公司 | Intelligent diagnosis guiding method and device for remote medical platform |
| CN114722977A (en) * | 2022-06-10 | 2022-07-08 | 四川大学 | Medical object classification method and device, electronic equipment and storage medium |
| CN114722977B (en) * | 2022-06-10 | 2022-09-02 | 四川大学 | Medical object classification method and device, electronic equipment and storage medium |
| CN116434954A (en) * | 2023-03-08 | 2023-07-14 | 绍兴珂西生物科技有限公司 | Evaluation method of clinical treatment effect |
| CN116434954B (en) * | 2023-03-08 | 2023-11-28 | 绍兴珂西生物科技有限公司 | Evaluation method of clinical treatment effect |
| CN116434954B8 (en) * | 2023-03-08 | 2023-12-22 | 绍兴珂西生物科技有限公司 | Evaluation method of clinical treatment effect |
| CN116386856B (en) * | 2023-06-05 | 2023-10-20 | 之江实验室 | Multi-label disease auxiliary diagnosis system based on doctor decision mode identification |
| CN116386856A (en) * | 2023-06-05 | 2023-07-04 | 之江实验室 | Multi-label disease auxiliary diagnosis system based on doctor decision mode identification |
| CN117558461A (en) * | 2024-01-12 | 2024-02-13 | 四川互慧软件有限公司 | Similar snake bite medical scheme selection method and device in different regions and electronic equipment |
| CN117558461B (en) * | 2024-01-12 | 2024-03-29 | 四川互慧软件有限公司 | Similar snake bite medical scheme selection method and device in different regions and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109411082B (en) | 2022-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109411082A (en) | A kind of Evaluation of Medical Quality and medical recommended method | |
| US20240203599A1 (en) | Method and system of for predicting disease risk based on multimodal fusion | |
| CN108648827B (en) | Cardiovascular and cerebrovascular disease risk prediction method and device | |
| CN109086805B (en) | Clustering method based on deep neural network and pairwise constraints | |
| WO2016192612A1 (en) | Method for analysing medical treatment data based on deep learning, and intelligent analyser thereof | |
| Chattopadhyay et al. | A Case‐Based Reasoning system for complex medical diagnosis | |
| WO2021120934A1 (en) | Convolutional neural network-based method for automatically grouping drgs | |
| Liu et al. | Deep learning based syndrome diagnosis of chronic gastritis | |
| Baker et al. | Continuous and automatic mortality risk prediction using vital signs in the intensive care unit: a hybrid neural network approach | |
| CN109378066A (en) | A kind of control method and control device for realizing disease forecasting based on feature vector | |
| CN108280149A (en) | A kind of doctor-patient dispute class case recommendation method based on various dimensions tag along sort | |
| CN109994216A (en) | A kind of ICD intelligent diagnostics coding method based on machine learning | |
| CN110838368A (en) | Robot active inquiry method based on traditional Chinese medicine clinical knowledge graph | |
| CN108511056A (en) | Therapeutic scheme based on patients with cerebral apoplexy similarity analysis recommends method and system | |
| CN117271804B (en) | Method, device, equipment and medium for generating common disease feature knowledge base | |
| CN113539460A (en) | Intelligent diagnosis guiding method and device for remote medical platform | |
| Johnson et al. | Encoding high-dimensional procedure codes for healthcare fraud detection | |
| CN114511759A (en) | Method and system for identifying categories and determining characteristics of skin state images | |
| WO2023217737A1 (en) | Health data enrichment for improved medical diagnostics | |
| CN109859813A (en) | A kind of entity modification word recognition method and device | |
| Chantamit-O-Pas et al. | A case-based reasoning framework for prediction of stroke | |
| Harnsomburana et al. | Computable visually observed phenotype ontological framework for plants | |
| CN112309519A (en) | Electronic medical record medication structured processing system based on multiple models | |
| CN113688854A (en) | Data processing method and device and computing equipment | |
| CN109036587A (en) | A kind of Community mental health services quality control system based on big data analysis |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Liu Xingwei Inventor after: Zhu Ke Inventor after: Liu Yang Inventor after: Chen Qiqi Inventor after: Liao Mingyang Inventor after: Mou Feng Inventor before: Liu Xingwei Inventor before: Liu Yang Inventor before: Chen Qiqi Inventor before: Liao Mingyang Inventor before: He Yi Inventor before: Mou Feng |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230301 Address after: Room 801, 85 Kefeng Road, Huangpu District, Guangzhou City, Guangdong Province Patentee after: Yami Technology (Guangzhou) Co.,Ltd. Address before: 610039, No. 999, Jin Zhou road, Jinniu District, Sichuan, Chengdu Patentee before: XIHUA University Effective date of registration: 20230301 Address after: 518000 1106, Building 8, Huali Business Center, Labor Community, Xixiang Street, Bao'an District, Shenzhen, Guangdong Province Patentee after: Shenzhen Juyue Medical Technology Co.,Ltd. Address before: Room 801, 85 Kefeng Road, Huangpu District, Guangzhou City, Guangdong Province Patentee before: Yami Technology (Guangzhou) Co.,Ltd. |