CN115050435A - Clinical trial data management system and method based on big data - Google Patents
Clinical trial data management system and method based on big data Download PDFInfo
- Publication number
- CN115050435A CN115050435A CN202210650994.6A CN202210650994A CN115050435A CN 115050435 A CN115050435 A CN 115050435A CN 202210650994 A CN202210650994 A CN 202210650994A CN 115050435 A CN115050435 A CN 115050435A
- Authority
- CN
- China
- Prior art keywords
- clinical
- subject
- data
- model
- value
- 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
Images
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
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/20—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for electronic clinical trials or questionnaires
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- 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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a clinical trial data management system and method based on big data, and belongs to the technical field of big data management. The system comprises the following modules: the system comprises a clinical test data management module, a deviation early warning module, a subject tracking and researching module and a scheduling follow-up module; the clinical test data management module is used for acquiring clinical research data and generating a big data model of the clinical test data; the deviation early warning module is used for outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port; the subject tracking investigation module is used for calculating an estimated value of the physical emergency of the subject; the scheduling follow-up module is used for scheduling a local medical institution to perform medical follow-up on the marked subjects. The invention can accurately expand the data of clinical tests, more accurately and variously analyze clinical symptoms, and simultaneously schedule the follow-up visits of the testees, thereby ensuring the medical conditions of the testees.
Description
Technical Field
The invention relates to the technical field of big data management, in particular to a clinical test data management system and method based on big data.
Background
With the continuous progress of clinical trial technology, the source data of clinical trials show the development trend of electronization. Currently, the source data of the clinical test in China is in a state of coexistence of paper and electronic forms. However, with the development of the times, the advantages of the electronic source data gradually emerge in the aspects of acquisition mode, traceability, quality standard, management authorization, file storage, safety guarantee and the like of the source data.
China 'quality management standard of clinical trials on drugs' stipulates that: "the medical record is used as the original file of clinical test and should be stored completely. The data in the case report table is from the original file and is consistent with the original file, and any observation and inspection results in the test should be timely, accurate, complete, normative and truly recorded in the case history and correctly filled in the case report table ". Therefore, in clinical trial data, the verification of the source data is an important step for ensuring the true and complete clinical trial data, is the basic responsibility of a clinical inspector, and is an indispensable quality control step in the process of transcribing or transcribing the source data from an original record.
Clinical symptoms recorded in clinical test data originate from a subject, but the subject cannot completely cover all clinical symptoms, and only can narrow the clinical symptoms within a certain interval range, and the interval range is used as a normal reaction of clinical medicines or tests, so that how to analyze the clinical symptoms more comprehensively and how to reduce the range of the clinical symptoms are the problems which are not solved at present; meanwhile, the follow-up treatment at the later stage of the subject is also a blank at present for analyzing the degree of severity and urgency of different clinical symptoms of the subject.
Disclosure of Invention
The present invention is directed to a clinical trial data management system and method based on big data, so as to solve the problems mentioned in the background art.
In order to solve the technical problems, the invention provides the following technical scheme:
a clinical trial data management method based on big data comprises the following steps:
s1, acquiring clinical research data and generating a big data model of clinical test data;
s2, acquiring a body state deviation value of the subject, constructing a deviation early warning threshold value, and outputting the information of the subject with the body state deviation value lower than the deviation early warning threshold value to an administrator port;
s3, the administrator port acquires the subject information data, marks the subject, and calculates the body emergency estimated value of the subject;
and S4, in the face of different body emergency estimated values of the testees, the administrator port sends scheduling information to schedule local medical institutions to carry out medical follow-up on the marked testees.
According to the above technical solution, the big data model of the clinical trial data includes:
acquiring clinical research data, wherein the clinical research data refers to historical data during a period of clinical research not yet tested, and generating a big data model of the clinical test data through the clinical research data, wherein the big data model of the clinical test data comprises clinical symptoms appearing in the clinical research data;
the big data model of the clinical test data further comprises: constructing a countermeasure network to correct the big data model of the clinical test data;
the countermeasure network comprises a generation model and a countermeasure model; the generation model is used for acquiring the characteristic vector of the clinical symptom which appears, and randomly selecting and fitting the characteristic vector into a new clinical symptom; the confrontation model is used for judging whether one clinical symptom belongs to clinical test data, setting probability to judge in a two-classification mode, and outputting 1 if the clinical symptom accords with the clinical test data; outputting 0 when the clinical symptoms do not accord with the clinical test data, and simultaneously outputting a probability value;
in the process of training the confrontation network, the method comprises the following steps:
s2-1, training an confrontation model, inputting clinical symptoms into the confrontation model, and judging whether the input clinical symptoms conform to the clinical test data or not by the confrontation model according to a big data model of the clinical test data and selecting the clinical symptoms conforming to the clinical test data;
s2-2, starting to train a generation model, wherein the generation model comprises a generation network, the generation network comprises a feature vector which randomly initializes a normally distributed clinical symptom, maps the feature vector to a higher dimension, fixes the parameters of the countermeasure model, optimizes the parameters of the generation model through a back propagation algorithm, and finally generates a clinical symptom similar to the clinical symptom which already appears, and can improve the probability that the generated clinical symptom which does not accord with the clinical test data is judged to accord with the clinical symptom of the clinical test data by the countermeasure model;
because clinical symptoms cannot be expressed by fixed values in clinical trials, the judgment is mainly performed through feature points, that is, in the process of judging the confrontation model, the judgment is usually within a range or a probability, and the confrontation network needs to make the range continuously narrow, so that the judgment result is more and more accurate. For example, a subject may have an allergic reaction after using a clinical drug, and the allergic reaction may be measured as a feature vector, for example, the allergic reaction is skin allergy eruption, and the area, density, size, itching, etc. of the eruption are all in a range, how to define the range, how to make the range smaller and smaller, that is, continuously learning through the anti-network of the present application, continuously adding the fitting of the feature vector, and finally obtaining more clinical symptom details to achieve better analysis effect.
S2-3, optimizing the confrontation model, calculating a loss function between the initialized clinical symptoms and the clinical symptoms generated by the optimized generation model through the confrontation model, and reducing the probability that the generated clinical symptoms which do not accord with the clinical test data are distinguished as the clinical symptoms which accord with the clinical test data by the optimized confrontation model;
s2-4, recording steps S2-2 to S2-3 as an iteration process, constructing iteration times, continuously repeating the steps S2-2 to S2-3 until the iteration times are reached, stopping repeating, and outputting a big data model of the corrected clinical test data;
the countermeasure network training process further comprises:
the generation model comprises a generation network A, and a clinical symptom similar to the clinical symptom already appeared but not belonging to the current clinical test data is generated by using the generation network A; similar distribution of clinical symptom data is recorded as P A (x; α) and the data distribution of the clinical symptoms which have appeared is P (x), wherein α represents a parameter for generating the network A and satisfies P A (x; α) has the highest similarity to P (x);
b characteristic points are selected from similar clinical symptom data distribution and are marked as a set { m } 1 、m 1 、…、m b Calculating likelihood data of b feature points according to the parameter alpha of the generated network A:
the maximum likelihood function estimate is:
L 1 ＝argmin α KL(P(x)‖P A (x；α))
wherein KL represents divergence and is used for measuring the similarity degree of the two probability distributions, and the smaller the numerical value is, the closer the two probability distributions are;
the generative model generates a network parameter alpha for continuous training, and a maximum likelihood function estimation value is searched to ensure that P is A The KL divergence between (x; alpha) and P (x) is minimal;
the confrontation model comprises:
the countermeasure model comprises a countermeasure network C, and an objective function of the countermeasure model is constructed as follows:
V(A，C)＝∫[P(x)lnC+P A (x；α)ln(1-C)]dx
wherein, V (A, C) represents an objective function of the confrontation model under the premise that the generated network is A;
because the confrontation model is a constant measure of P A (x; alpha) and P (x) to reduce the resulting clinical symptoms that do not meet clinical trial data, and to optimize the challenge modelSince the probability of the clinical symptom corresponding to the clinical trial data is discriminated, the larger V (a, C) is, the better the discrimination effect is, and the maximum value of V (a, C) is calculated:
wherein maxV (A, C) represents the maximum value of V (A, C);
in the countermeasure network, acquiring a countermeasure model, and then fixing the countermeasure network of the countermeasure model to obtain an initial countermeasure model; then, on the basis of the initial confrontation model, training a generation model to ensure that the KL divergence of probability distribution between the generation network and real data is minimum, and obtaining a final optimization mode N:
N＝argmin A max C V(A，C)
and obtaining the big data model of the clinical test data after the optimization mode as a new big data model of the clinical test data, and outputting the new big data model as final output.
In the above technical solution, the accuracy of the determination can be improved by continuously training, for example, an initial clinical symptom I has a probability of falling on the clinical trial data of 80, the generation network a generates a clinical symptom similar to the clinical symptom I that has already appeared, and has a probability of falling on the clinical trial data of 52, because 50 is set as a probability point, it is a clinical symptom belonging to the current clinical trial data, but actually, a loss function between the clinical symptom I and the clinical symptom similar to the clinical symptom I is not considered, after the processing by using the optimization mode N, the accuracy of the determination is further improved, and the calculated clinical symptom similar to the clinical symptom I may fall on a probability of falling on the clinical trial data of less than 50, and then the system will pick out the clinical symptom.
According to the above technical solution, the body state deviation value of the subject includes:
acquiring a big data model of the clinical test data output in the step S1, and acquiring clinical symptoms of the subject;
inputting the data into a big data model of clinical trial data, and acquiring the probability of the data falling into the clinical trial data;
calculating a body state deviation value of the subject:
Q 1 ＝q 1 -q 0
wherein Q is 1 A deviation value representing the physical state of the subject; q. q.s 1 Represents the probability of the subject's clinical symptoms falling into the clinical trial data; q. q.s 0 A probability threshold representing a clinical symptom for the subject;
and constructing a deviation early warning threshold value, and outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port.
According to the above technical solution, the calculating the subject body emergency estimate value includes:
the administrator port acquires subject information data, wherein the subject information data comprises subject clinical symptom time data and subject location point data;
calculating a physical emergency estimate for the subject:
y 0 ＝k 1 *Q 1
wherein, y 0 Representing a physical emergency estimate of the subject; k is a radical of 1 A growth coefficient representing a physical emergency of the subject, satisfying k 1 >0; the lower the subject's physical emergency estimate, the more critical the subject's condition is represented.
In the technical scheme, after the clinical symptoms of the subject are obtained, the clinical symptoms are input into a big data model of the corrected clinical test data, the probability of falling into the clinical test data can be obtained, the probability and the probability difference value are calculated, the greater the probability of falling into the clinical test data is, the clinical symptoms of the current subject exist or are similar in clinical research or historical data, the current subject belongs to a controllable range, and the influence on the subject is prepared in a predetermined way; the smaller the probability of falling into the clinical trial data, the more likely it is that the clinical symptoms of the current subject are absent or similarly low in the clinical study or historical data, which indicates that the subject is in a critical state when a completely new clinical symptom is present, possibly due to personal constitution or other reasons.
According to the above technical solution, the sending of the scheduling information by the administrator port includes:
acquiring a medical institution of a place where a subject is located; acquiring the time of each medical institution reaching the position of the subject;
acquiring medical resources and historical diagnosis and treatment data of each medical institution, and generating average diagnosis and treatment duration of each medical institution:
wherein, T g Represents the average diagnosis and treatment time of the medical institution g; j represents a serial number;
beta represents the number of people participating in diagnosis and treatment; t is t j Representing the diagnosis and treatment time of the jth patient;
constructing a scheduling model:
setting a subject physical emergency estimate threshold y 1 (ii) a At y 0 >y 1 The scheduling influence value U of the medical institution g g1 ：
U g1 ＝y 0 *k 2 +(T g *m g +t e，g )*k 3 +S g *k 4
Wherein k is 2 Is represented by y 0 >y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g1 The influence coefficient value of (a); m is g A waiting patient on behalf of the current medical institution g; t is t e，g
Represents the time from the medical institution g to the location of the subject; k is a radical of 3 Is represented by y 0 >y 1 Time-to-time scheduling impact U for medical facility g g1 The influence coefficient value of (a); s g Manpower and material resources representing the cost of the distance from the medical institution g to the location of the subject; k is a radical of formula 4 Is represented by y 0 >y 1 The scheduling influence value U of the distance spent on the medical institution g g1 Value of influence coefficient of；
At y 0 ≤y 1 The scheduling influence value U of the medical institution g g2 ：
U g2 ＝y 0 *k 5 +t e，g *k 3
Wherein k is 5 Is represented by y 0 ≤y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g2 The influence coefficient value of (a);
respectively calculating the scheduling influence value of each medical institution;
and selecting the medical institution with the lowest scheduling influence value, sending scheduling information through the administrator port, and scheduling the medical institution to perform medical follow-up on the marked testees.
A big data based clinical trial data management system, the system comprising the following modules: the system comprises a clinical test data management module, a deviation early warning module, a subject tracking and researching module and a scheduling follow-up module;
the clinical test data management module is used for acquiring clinical research data and generating a big data model of the clinical test data; the deviation early warning module is used for acquiring a body state deviation value of a subject, constructing a deviation early warning threshold value and outputting the information of the subject with the body state deviation value lower than the deviation early warning threshold value to an administrator port; the subject tracking and researching module is used for acquiring subject information data, collecting clinical symptoms of a subject, marking the subject and calculating an estimated value of the physical emergency of the subject; the scheduling follow-up module is used for sending scheduling information by the administrator port according to the condition of a local medical institution when facing different body emergency estimated values of the testees, and scheduling the local medical institution to perform medical follow-up on the marked testees;
the output end of the clinical test data management module is connected with the input end of the deviation early warning module; the output end of the deviation early warning module is connected with the input end of the subject tracking and researching module; the output end of the subject tracking investigation module is connected with the input end of the dispatching follow-up module.
According to the technical scheme, the clinical test data management module comprises a clinical research data analysis submodule and a model correction submodule;
the clinical research data analysis submodule is used for acquiring clinical research data and processing the clinical research data; the model correction submodule is used for generating a big data model of clinical test data according to the clinical research data and correcting the big data model;
the output end of the clinical research data analysis submodule is connected with the input end of the model correction submodule; and the output end of the model correction submodule is connected with the input end of the deviation early warning module.
According to the technical scheme, the deviation early warning module comprises a body state deviation value calculation operator module of the subject and a marking early warning sub-module;
the body state deviation value operator module of the subject is used for calculating the body state deviation value of the subject:
Q 1 ＝q 1 -q 0
wherein Q is 1 A deviation value representing the physical state of the subject; q. q.s 1 Represents the probability of the subject's clinical symptoms falling into the clinical trial data; q. q.s 0 A probability threshold representing a clinical symptom for the subject;
the mark early warning sub-module is used for constructing a deviation early warning threshold value and outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port;
the output end of the operator module for calculating the body state deviation value of the subject is connected with the input end of the mark early warning sub-module; the output end of the mark early warning sub-module is connected with the input end of the subject tracking and researching module.
According to the technical scheme, the subject tracking and research module comprises a subject clinical symptom acquisition sub-module and a subject body emergency situation estimation value calculation sub-module;
the subject clinical symptom acquisition sub-module is used for acquiring subject information data, acquiring clinical symptoms of the subject and marking the subject; the subject body emergency situation estimation value calculation sub-module is used for calculating a subject body emergency situation estimation value according to clinical symptoms of the subject;
the output end of the subject clinical symptom acquisition submodule is connected with the input end of the subject body emergency situation estimation value calculation submodule; and the output end of the body emergency situation estimation value calculation submodule of the subject is connected with the input end of the dispatching follow-up module.
According to the technical scheme, the scheduling follow-up module comprises a data analysis submodule and an information scheduling submodule;
the data analysis submodule is used for calculating the scheduling influence value of each medical institution by combining the conditions of local medical institutions when different body emergency estimated values of the testees face;
the calculation of the scheduling impact value for the medical institution comprises:
constructing a scheduling model:
setting a subject physical emergency estimate threshold y 1 (ii) a At y 0 >y 1 The scheduling influence value U of the medical institution g g1 ：
U g1 ＝y 0 *k 2 +(T g *m g +t e，g )*k 3 +S g *k 4
Wherein k is 2 Is represented by y 0 >y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g1 The influence coefficient value of (a); m is g A waiting patient on behalf of the current medical institution g; t is t e，g
Represents the time from the medical institution g to the location of the subject; k is a radical of 3 Is represented by y 0 >y 1 Time-to-time scheduling impact U for medical facility g g1 The influence coefficient value of (a); s g Manpower and material resources representing the cost of the distance from the medical institution g to the location of the subject; k is a radical of 4 Is represented by y 0 >y 1 The scheduling influence value U of the distance spent on the medical institution g g1 The influence coefficient value of (a); y is 0 Representing a physical emergency estimate of the subject; t is g Represents the average diagnosis and treatment time of the medical institution g;
at y 0 ≤y 1 The scheduling influence value U of the medical institution g g2 ：
U g2 ＝y 0 *k 5 +t e，g *k 3
Wherein k is 5 Is represented by y 0 ≤y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g2 The influence coefficient value of (a);
respectively calculating the scheduling influence value of each medical institution;
the information scheduling submodule is used for selecting the medical institution with the lowest scheduling influence value, and the administrator port sends scheduling information to schedule the medical institution to perform medical follow-up visit on the marked subjects.
Compared with the prior art, the invention has the following beneficial effects: according to the invention, a clinical test data management module is used for acquiring clinical research data and generating a big data model of the clinical test data; acquiring a body state deviation value of a subject by using a deviation early warning module, constructing a deviation early warning threshold value, and outputting the information of the subject with the body state deviation value lower than the deviation early warning threshold value to an administrator port; acquiring information data of the testee by using a testee tracking and researching module, acquiring clinical symptoms of the testee, marking the testee, and simultaneously calculating an estimated value of the physical emergency of the testee; when a scheduling follow-up module is used for facing different body emergency estimated values of the testees, scheduling information is sent by an administrator port according to the condition of a local medical institution, and the local medical institution is scheduled to perform medical follow-up on the marked testees; the invention can continuously and intelligently learn and evolve in the aspect of clinical data management, utilizes the antagonistic network thought, and continuously increases the characteristics of clinical symptoms, so that the control of the clinical symptoms in clinical test data is more accurate, and meanwhile, the contradiction between patient privacy disclosure caused by manual copying of symptoms and doctors and patients caused by manual errors is avoided; meanwhile, the medical treatment system is also provided with a scheduling model, so that more attention can be paid when the testee has unexpected clinical symptoms, and a local medical institution can be scheduled for follow-up visit in an emergency, so that the medical conditions of the testee can be preferentially guaranteed.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a clinical trial data management system and method based on big data according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, in the present embodiment:
acquiring clinical research data, and generating a big data model of clinical test data;
the clinical research data refers to historical data during the period of clinical research not yet tested, and a big data model of the clinical research data is generated through the clinical research data, and the big data model of the clinical research data comprises clinical symptoms appearing in the clinical research data;
the big data model of the clinical test data further comprises: constructing a countermeasure network to correct the big data model of the clinical test data;
the countermeasure network comprises a generation model and a countermeasure model; the generation model is used for acquiring the characteristic vector of the clinical symptom which appears, and randomly selecting and fitting the characteristic vector into a new clinical symptom; the confrontation model is used for judging whether one clinical symptom belongs to clinical test data, setting probability and judging in a two-classification mode, and outputting 1 if the clinical symptom accords with the clinical test data; outputting 0 when the clinical symptoms do not accord with the clinical test data, and simultaneously outputting a probability value; for example, the probability is set to 50, if the probability exceeds 50, the probability is 1, and the output probability value is 100%;
in the process of training the confrontation network, the method comprises the following steps:
training an confrontation model, inputting clinical symptoms into the confrontation model, and judging by the confrontation model to select the clinical symptoms which accord with clinical test data;
training a generation model, wherein the generation model comprises a generation network, the generation network comprises a feature vector for randomly initializing a normally distributed clinical symptom, mapping the feature vector to a higher dimension, fixing parameters of the countermeasure model, optimizing parameters of the generation model through a back propagation algorithm, and finally generating a clinical symptom similar to the presented clinical symptom, and the probability that the generated clinical symptom which does not accord with the clinical test data is judged to accord with the clinical symptom of the clinical test data by the countermeasure model can be improved;
optimizing the countermeasure model, calculating a loss function between the initialized clinical symptoms and the clinical symptoms generated by the optimized generation model through the countermeasure model, and reducing the probability that the generated clinical symptoms which do not accord with the clinical test data are judged to accord with the clinical symptoms of the clinical test data by the optimized countermeasure model;
constructing a big data model of the clinical test data after the iteration times are n times, continuously repeating the steps, continuously generating and optimizing until the iteration times are reached, terminating, and outputting the corrected clinical test data;
the countermeasure network training process further comprises:
the generation model comprises a generation network A, and a clinical symptom similar to the presented clinical symptom but not belonging to the current clinical test data is generated by using the generation network A; similar distribution of clinical symptom data is recorded as P A (x; α) and the data distribution of the clinical symptoms which have appeared is P (x), wherein α represents a parameter for generating the network A and satisfies P A (x; α) has the highest similarity to P (x);
b from the similar clinical symptom data distributionCharacteristic points, denoted as set { m } 1 、m 1 、…、m b Calculating likelihood data of b feature points according to the parameter alpha of the generated network A:
the maximum likelihood function estimate is:
L 1 ＝argmin α KL(P(x)‖P A (x；α))
wherein KL represents divergence and is used for measuring the similarity degree of the two probability distributions, and the smaller the numerical value is, the closer the two probability distributions are;
the generative model generates a network parameter alpha for continuous training, and a maximum likelihood function estimation value is searched to ensure that P is A The KL divergence between (x; alpha) and P (x) is minimal;
the confrontation model comprises:
the countermeasure model comprises a countermeasure network C, and an objective function of the countermeasure model is constructed as follows:
V(A，C)＝∫[P(x)lnC+P A (x；α)ln(1-C)]dx
wherein, V (A, C) represents an objective function of the confrontation model under the premise that the generated network is A;
because the confrontation model is a constant measure of P A (x; α) and P (x) to reduce the probability that the resulting clinical symptom not meeting the clinical trial data is judged by the optimized countermeasure model as meeting the clinical symptom of the clinical trial data, so that the larger V (A, C), the better the judgment, the maximum value of V (A, C) is calculated:
wherein maxV (A, C) represents the maximum value of V (A, C);
in the countermeasure network, acquiring a countermeasure model, and then fixing the countermeasure network of the countermeasure model to obtain an initial countermeasure model; then, on the basis of the initial confrontation model, training a generation model to ensure that the KL divergence of probability distribution between the generation network and real data is minimum, and obtaining a final optimization mode N:
N＝argmin A max C V(A，C)
and obtaining the big data model of the clinical test data after the optimization mode as a new big data model of the clinical test data, and outputting the new big data model as final output.
Calculating a body state deviation value of the subject:
obtaining clinical symptoms of a subject;
inputting the data into a big data model of the corrected clinical trial data, and acquiring the probability of the data falling into the clinical trial data;
calculating a body state deviation value of the subject:
Q 1 ＝q 1 -q 0
wherein Q is 1 A deviation value representing the physical state of the subject; q. q.s 1 Represents the probability of the subject's clinical symptoms falling into the clinical trial data; q. q.s 0 A probability threshold representing a clinical symptom for the subject;
and constructing a deviation early warning threshold value, and outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port.
Calculating a physical emergency estimate for the subject:
the administrator port acquires subject information data, wherein the subject information data comprises subject clinical symptom time data and subject location point data;
calculating a subject physical emergency estimate:
y 0 ＝k 1 *Q 1
wherein, y 0 Representing a physical emergency estimate of the subject; k is a radical of 1 A growth coefficient representing a physical emergency of the subject, satisfying k 1 >0; the lower the subject's physical emergency estimate, the more critical the subject's condition is represented.
Acquiring a medical institution of a place where a subject is located; acquiring the time of each medical institution reaching the position of the subject;
acquiring medical resources and historical diagnosis and treatment data of each medical institution, and generating average diagnosis and treatment duration of each medical institution:
wherein, T g Represents the average diagnosis and treatment time of the medical institution g; j represents a serial number;
beta represents the number of people participating in diagnosis and treatment; t is t j Representing the diagnosis and treatment time of the jth patient;
constructing a scheduling model:
setting a subject physical emergency estimate threshold y 1 (ii) a At y 0 >y 1 The scheduling influence value U of the medical institution g g1 ：
U g1 ＝y 0 *k 2 +(T g *m g +t e，g )*k 3 +S g *k 4
Wherein k is 2 Is represented by y 0 >y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g1 The influence coefficient value of (a); m is g A waiting patient on behalf of the current medical institution g; t is t e，g
Represents the time from the medical institution g to the location of the subject; k is a radical of 3 Is represented by y 0 >y 1 Time-to-time scheduling impact U for medical facility g g1 The influence coefficient value of (a); s g Manpower and material resources representing the cost of the distance from the medical institution g to the location of the subject; k is a radical of 4 Is represented by y 0 >y 1 The scheduling influence value U of the distance spent on the medical institution g g1 The influence coefficient value of (a);
at y 0 ≤y 1 The scheduling influence value U of the medical institution g g2 ：
U g2 ＝y 0 *k 5 +t e，g *k 3
Wherein k is 5 Is represented by y 0 ≤y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g2 The influence coefficient value of (a);
respectively calculating the scheduling influence value of each medical institution;
and selecting the medical institution with the lowest scheduling influence value, sending scheduling information through the administrator port, and scheduling the medical institution to perform medical follow-up on the marked testees.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A clinical trial data management method based on big data is characterized in that: the method comprises the following steps:
s1, acquiring clinical research data and generating a big data model of clinical test data;
s2, acquiring a body state deviation value of the subject, constructing a deviation early warning threshold value, and outputting the information of the subject with the body state deviation value lower than the deviation early warning threshold value to an administrator port;
s3, the administrator port acquires the subject information data, marks the subject, and calculates the body emergency estimated value of the subject;
and S4, in the face of different body emergency estimated values of the testees, the administrator port sends scheduling information to schedule local medical institutions to carry out medical follow-up on the marked testees.
2. The big data based clinical trial data management method according to claim 1, wherein: the big data model of the clinical trial data comprises:
acquiring clinical research data which refers to historical data during a period of clinical research not yet tested, and generating a big data model of the clinical test data through the clinical research data, wherein the big data model of the clinical test data comprises clinical symptoms appearing in the clinical research data;
the big data model of the clinical test data further comprises: constructing a countermeasure network to correct the big data model of the clinical test data;
the countermeasure network comprises a generation model and a countermeasure model; the generation model is used for acquiring the characteristic vector of the clinical symptom which appears, and randomly selecting and fitting the characteristic vector into a new clinical symptom; the confrontation model is used for judging whether one clinical symptom belongs to clinical test data, setting probability to judge in a two-classification mode, and outputting 1 if the clinical symptom accords with the clinical test data; outputting 0 when the clinical symptoms do not accord with the clinical test data, and simultaneously outputting a probability value;
in the process of training the confrontation network, the method comprises the following steps:
s2-1, training an confrontation model, inputting clinical symptoms into the confrontation model, and judging whether the input clinical symptoms conform to the clinical test data or not by the confrontation model according to a big data model of the clinical test data and selecting the clinical symptoms conforming to the clinical test data;
s2-2, starting to train a generation model, wherein the generation model comprises a generation network, the generation network comprises a feature vector which randomly initializes a normally distributed clinical symptom, maps the feature vector to a higher dimension, fixes the parameters of the countermeasure model, optimizes the parameters of the generation model through a back propagation algorithm, and finally generates a clinical symptom similar to the clinical symptom which already appears, and can improve the probability that the generated clinical symptom which does not accord with the clinical test data is judged to accord with the clinical symptom of the clinical test data by the countermeasure model;
s2-3, optimizing the confrontation model, calculating a loss function between the initialized clinical symptoms and the clinical symptoms generated by the optimized generative model through the confrontation model, and reducing the probability that the generated clinical symptoms which do not accord with the clinical test data are judged to accord with the clinical symptoms of the clinical test data by the optimized confrontation model;
s2-4, recording steps S2-2 to S2-3 as an iteration process, constructing iteration times, continuously repeating the steps S2-2 to S2-3 until the iteration times are reached, stopping repeating, and outputting a big data model of the corrected clinical test data;
the countermeasure network training process further comprises:
the generation model comprises a generation network A, and a clinical symptom similar to the presented clinical symptom but not belonging to the current clinical test data is generated by using the generation network A; similar distribution of clinical symptom data is recorded as P A (x; α) and the data distribution of the clinical symptoms which have appeared is P (x), wherein α represents a parameter for generating the network A and satisfies P A (x; α) has the highest similarity to P (x);
b characteristic points are selected from similar clinical symptom data distribution and are marked as a set { m } 1 、m 1 、…、m b Calculating likelihood data of b feature points according to the parameter alpha of the generated network A:
the maximum likelihood function estimate is:
L 1 ＝argmin α KL(P(x)||P A (x；α))
wherein KL represents divergence and is used for measuring the similarity degree of the two probability distributions, and the smaller the numerical value is, the closer the two probability distributions are;
the generative model generates a network parameter alpha for continuous training, and a maximum likelihood function estimation value is searched to ensure that P is A The KL divergence between (x; alpha) and P (x) is minimal;
the confrontation model comprises:
the countermeasure model comprises a countermeasure network C, and an objective function of the countermeasure model is constructed as follows:
V(A，C)＝∫[P(x)lnC+P A (x；α)ln(1-C)]dx
wherein, V (A, C) represents an objective function of the confrontation model under the premise that the generated network is A;
because the confrontation model is a constant measure of P A (x; α) and P (x) to reduce the probability that the resulting clinical symptom not meeting the clinical trial data is judged by the optimized countermeasure model as meeting the clinical symptom of the clinical trial data, so that the larger V (A, C), the better the judgment, the maximum value of V (A, C) is calculated:
wherein maxV (A, C) represents the maximum value of V (A, C);
in the countermeasure network, acquiring a countermeasure model, and then fixing the countermeasure network of the countermeasure model to obtain an initial countermeasure model; then, on the basis of the initial confrontation model, training a generation model to ensure that the KL divergence of probability distribution between the generation network and real data is minimum, and obtaining a final optimization mode N:
N＝argmin A max C V(A,C)
and obtaining the big data model of the clinical test data after the optimization mode as a new big data model of the clinical test data, and outputting the new big data model as final output.
3. The big data based clinical trial data management method of claim, wherein: the subject's body state deviation value comprises:
acquiring a big data model of the clinical test data output in the step S1, and acquiring clinical symptoms of the subject;
inputting the data into a big data model of clinical trial data, and acquiring the probability of the data falling into the clinical trial data;
calculating a body state deviation value of the subject:
Q 1 ＝q 1 -q 0
wherein Q is 1 A deviation value representing the physical state of the subject; q. q.s 1 Represents the probability of the subject's clinical symptoms falling into the clinical trial data; q. q.s 0 A probability threshold representing a clinical symptom for the subject;
and constructing a deviation early warning threshold value, and outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port.
4. The big data based clinical trial data management method of claim 3, wherein: the calculating a subject physical emergency estimate comprises:
the administrator port acquires subject information data, wherein the subject information data comprises subject clinical symptom time data and subject location point data;
calculating a subject physical emergency estimate:
y 0 ＝k 1 *Q 1
wherein, y 0 Representing a physical emergency estimate of the subject; k is a radical of 1 A growth coefficient representing a physical emergency of the subject, satisfying k 1 >0; the lower the subject's physical emergency estimate, the more critical the subject's condition is represented.
5. The big-data based clinical trial data management method according to claim 4, wherein: the transmitting of the scheduling information by the administrator port includes:
acquiring a medical institution of a place where a subject is located; acquiring the time of each medical institution reaching the position of the subject;
acquiring medical resources and historical diagnosis and treatment data of each medical institution, and generating average diagnosis and treatment duration of each medical institution:
wherein, T g Represents the average diagnosis and treatment time of the medical institution g; j represents a serial number; beta represents the number of people participating in diagnosis and treatment; t is t j Representing the diagnosis and treatment time of the jth patient;
constructing a scheduling model:
setting a subject physical emergency estimate threshold y 1 (ii) a At y 0 ＞y 1 The scheduling influence value U of the medical institution g g1 ：
U g1 ＝y 0 *k 2 +(T g *m g +t e，g )*k 3 +S g *k 4
Wherein k is 2 Is represented by y 0 ＞y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g1 The influence coefficient value of (a); m is g A waiting patient on behalf of the current medical institution g; t is t e，g Represents the time from the medical institution g to the location of the subject; k is a radical of 3 Is represented by y 0 ＞y 1 Time-to-time scheduling impact U for medical facility g g1 The influence coefficient value of (a); s g Manpower and material resources representing the cost of the distance from the medical institution g to the location of the subject; k is a radical of 4 Is represented by y 0 ＞y 1 The scheduling influence value U of the distance spent on the medical institution g g1 The influence coefficient value of (a);
at y 0 ≤y 1 The scheduling influence value U of the medical institution g g2 ：
U g2 ＝y 0 *k 5 +t e，g ，k 3
Wherein k is 5 Is represented by y 0 ≤y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g2 The influence coefficient value of (a);
respectively calculating the scheduling influence value of each medical institution;
and selecting the medical institution with the lowest scheduling influence value, sending scheduling information through the administrator port, and scheduling the medical institution to perform medical follow-up on the marked testees.
6. A clinical trial data management system based on big data, characterized by: the system comprises the following modules: the system comprises a clinical test data management module, a deviation early warning module, a subject tracking and researching module and a scheduling follow-up module;
the clinical test data management module is used for acquiring clinical research data and generating a big data model of the clinical test data; the deviation early warning module is used for acquiring a body state deviation value of a subject, constructing a deviation early warning threshold value and outputting the information of the subject with the body state deviation value lower than the deviation early warning threshold value to an administrator port; the subject tracking and researching module is used for acquiring subject information data, collecting clinical symptoms of a subject, marking the subject and calculating an estimated value of the physical emergency of the subject; the scheduling follow-up module is used for sending scheduling information by the administrator port according to the condition of a local medical institution when facing different body emergency estimated values of the testees, and scheduling the local medical institution to perform medical follow-up on the marked testees;
the output end of the clinical test data management module is connected with the input end of the deviation early warning module; the output end of the deviation early warning module is connected with the input end of the subject tracking and researching module; the output end of the subject tracking investigation module is connected with the input end of the dispatching follow-up module.
7. The big-data based clinical trial data management system according to claim 6, wherein: the clinical test data management module comprises a clinical research data analysis submodule and a model correction submodule;
the clinical research data analysis submodule is used for acquiring clinical research data and processing the clinical research data; the model correction submodule is used for generating a big data model of clinical test data according to the clinical research data and correcting the big data model;
the output end of the clinical research data analysis submodule is connected with the input end of the model correction submodule; and the output end of the model correction submodule is connected with the input end of the deviation early warning module.
8. The big-data based clinical trial data management system according to claim 6, wherein: the deviation early warning module comprises a body state deviation value calculating operator module and a mark early warning sub-module of the subject;
the body state deviation value operator module of the subject is used for calculating the body state deviation value of the subject:
Q 1 ＝q 1 -q 0
wherein Q is 1 A deviation value representing the physical state of the subject; q. q.s 1 Represents the probability of the subject's clinical symptoms falling into the clinical trial data; q. q.s 0 A probability threshold representing a clinical symptom for the subject;
the mark early warning sub-module is used for constructing a deviation early warning threshold value and outputting the information of the subjects with the body state deviation value lower than the deviation early warning threshold value to an administrator port;
the output end of the operator module for calculating the body state deviation value of the subject is connected with the input end of the mark early warning sub-module; the output end of the mark early warning sub-module is connected with the input end of the testee tracking and researching module.
9. The big-data based clinical trial data management system according to claim 6, wherein: the subject tracking and researching module comprises a subject clinical symptom acquisition sub-module and a subject body emergency estimated value calculation sub-module;
the subject clinical symptom acquisition sub-module is used for acquiring subject information data, acquiring clinical symptoms of the subject and marking the subject; the subject body emergency situation estimation value calculation sub-module is used for calculating a subject body emergency situation estimation value according to clinical symptoms of a subject;
the output end of the subject clinical symptom acquisition submodule is connected with the input end of the subject body emergency situation estimation value calculation submodule; and the output end of the body emergency situation estimation value calculation submodule of the subject is connected with the input end of the dispatching follow-up module.
10. The big-data based clinical trial data management system according to claim 6, wherein: the scheduling follow-up module comprises a data analysis submodule and an information scheduling submodule;
the data analysis submodule is used for calculating the scheduling influence value of each medical institution by combining the conditions of local medical institutions when different body emergency estimated values of the testees face;
the calculation of the scheduling impact value for the medical institution comprises:
constructing a scheduling model:
setting a subject physical emergency estimate threshold y 1 (ii) a At y 0 ＞y 1 The scheduling influence value U of the medical institution g g1 ：
U g1 ＝y 0 *k 2 +(T g *m g +t e，g )*k 3 +S g *k 4
Wherein k is 2 Is represented by y 0 ＞y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g1 The influence coefficient value of (a); m is g A waiting patient on behalf of the current medical institution g; t is t e，g Represents the time from the medical institution g to the location of the subject; k is a radical of 3 Is represented by y 0 ＞y 1 Time-to-time medical institutiong scheduling impact value U g1 The influence coefficient value of (a); s g Manpower and material resources representing the cost of the distance from the medical institution g to the location of the subject; k is a radical of 4 Is represented by y 0 ＞y 1 The scheduling influence value U of the distance spent on the medical institution g g1 The influence coefficient value of (a); y is 0 Representing a physical emergency estimate of the subject; t is g Represents the average diagnosis and treatment time of the medical institution g;
at y 0 ≤y 1 The scheduling influence value U of the medical institution g g2 ：
U g2 ＝y 0 *k 5 +t e，g *k 3
Wherein k is 5 Is represented by y 0 ≤y 1 The estimated value of the physical emergency of the subject has a scheduling influence U on the medical institution g g2 The influence coefficient value of (a);
respectively calculating the scheduling influence value of each medical institution;
the information scheduling submodule is used for selecting the medical institution with the lowest scheduling influence value, and the administrator port sends scheduling information to schedule the medical institution to perform medical follow-up visit on the marked subjects.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210650994.6A CN115050435B (en) | 2022-06-09 | 2022-06-09 | Clinical trial data management system and method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210650994.6A CN115050435B (en) | 2022-06-09 | 2022-06-09 | Clinical trial data management system and method based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115050435A true CN115050435A (en) | 2022-09-13 |
| CN115050435B CN115050435B (en) | 2023-04-07 |
Family
ID=83162272
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210650994.6A Active CN115050435B (en) | 2022-06-09 | 2022-06-09 | Clinical trial data management system and method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115050435B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116072305A (en) * | 2023-02-17 | 2023-05-05 | 北京中兴正远科技有限公司 | Clinical trial data acquisition system |
| CN116631551A (en) * | 2023-07-21 | 2023-08-22 | 山东博达医疗用品股份有限公司 | Data management method based on black needle clinical treatment |
| CN116864050A (en) * | 2023-05-26 | 2023-10-10 | 中国人民解放军总医院 | Clinical trial quality control method and equipment for scheme deviation semi-quantitative evaluation |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108830034B (en) * | 2018-05-25 | 2021-07-20 | 陈育聪 | Medicine clinical research and development information processing system and method |
| US20240006052A1 (en) * | 2020-09-25 | 2024-01-04 | Cornell University | System implementing generative adversarial network adapted to prediction in behavioral and/or physiological contexts |
| CN113707248A (en) * | 2021-08-31 | 2021-11-26 | 郑州大学第一附属医院 | Clinical trial subject examination and inspection process management system, method and computer storage medium |
| CN114300108A (en) * | 2021-12-30 | 2022-04-08 | 浙江太美医疗科技股份有限公司 | Clinical test item implementer recommendation method, device, equipment and storage medium |
-
2022
- 2022-06-09 CN CN202210650994.6A patent/CN115050435B/en active Active
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116072305A (en) * | 2023-02-17 | 2023-05-05 | 北京中兴正远科技有限公司 | Clinical trial data acquisition system |
| CN116072305B (en) * | 2023-02-17 | 2023-09-29 | 北京中兴正远科技有限公司 | Clinical trial data acquisition system |
| CN116864050A (en) * | 2023-05-26 | 2023-10-10 | 中国人民解放军总医院 | Clinical trial quality control method and equipment for scheme deviation semi-quantitative evaluation |
| CN116631551A (en) * | 2023-07-21 | 2023-08-22 | 山东博达医疗用品股份有限公司 | Data management method based on black needle clinical treatment |
| CN116631551B (en) * | 2023-07-21 | 2023-09-26 | 山东博达医疗用品股份有限公司 | Data management method based on black needle clinical treatment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115050435B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115050435B (en) | Clinical trial data management system and method based on big data | |
| CN102985007B (en) | Health care device monitoring arrangement | |
| CN110706786B (en) | Non-contact intelligent psychological parameter analysis and evaluation system | |
| KR20170061222A (en) | The method for prediction health data value through generation of health data pattern and the apparatus thereof | |
| US11529105B2 (en) | Digital twin updating | |
| CN116246768B (en) | MRI image inspection intelligent analysis management system based on artificial intelligence | |
| CN116258726A (en) | Temporal-mandibular joint MRI image important structure segmentation method based on deep learning | |
| CN115517681A (en) | Method and system for monitoring mood fluctuation and evaluating emotional disorder state of MD (MD) patient | |
| CN110192860B (en) | Brain imaging intelligent test analysis method and system for network information cognition | |
| Feng et al. | Using eye aspect ratio to enhance fast and objective assessment of facial paralysis | |
| CN117617921A (en) | Intelligent blood pressure monitoring system and method based on Internet of things | |
| CN117292836A (en) | Body measurement score monitoring method and system | |
| Oleson | Bayesian credible intervals for binomial proportions in a single patient trial | |
| CN115472312A (en) | Medical health management system and method based on medical defense fusion service | |
| CN110859599B (en) | Automatic cognitive function screening system for cerebrovascular disease nerve injury patients | |
| Koşan et al. | Developing an item bank for progress tests and application of computerized adaptive testing by simulation in medical education | |
| CN116196034B (en) | Ultrasonic imaging method and system | |
| CN112382385A (en) | Aura-free migraine auxiliary diagnosis algorithm based on 3D convolutional neural network | |
| CN112086193A (en) | Face recognition health prediction system and method based on Internet of things | |
| CN116092665B (en) | Ophthalmic screening diagnosis and treatment system based on artificial intelligence | |
| CN115393353B (en) | Medical image data quality monitoring and management method and system | |
| Aswini et al. | For Effective, Earlier and Simplified Diagnosis of Retinopathy of Prematurity (RoP), a Probe through Digital Image Processing Algorithm in B-Scan | |
| CN117951190B (en) | Human body index abnormal data processing method and system based on artificial intelligence | |
| CN118098478A (en) | Physical examination information analysis management system | |
| CN114550860B (en) | Hospitalizing satisfaction evaluation method based on process data and intelligent network model |
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 |