CN110033863A - Applied to the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems, device - Google Patents

Applied to the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems, device Download PDF

Info

Publication number
CN110033863A
CN110033863A CN201910328888.4A CN201910328888A CN110033863A CN 110033863 A CN110033863 A CN 110033863A CN 201910328888 A CN201910328888 A CN 201910328888A CN 110033863 A CN110033863 A CN 110033863A
Authority
CN
China
Prior art keywords
disease
target
medical diagnosis
subgraph
result
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
Application number
CN201910328888.4A
Other languages
Chinese (zh)
Other versions
CN110033863B (en
Inventor
刘明录
吴及
吕萍
胡飞
李伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
iFlytek Co Ltd
Original Assignee
Tsinghua University
iFlytek Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tsinghua University, iFlytek Co Ltd filed Critical Tsinghua University
Priority to CN201910328888.4A priority Critical patent/CN110033863B/en
Publication of CN110033863A publication Critical patent/CN110033863A/en
Application granted granted Critical
Publication of CN110033863B publication Critical patent/CN110033863B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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
    • G16H50/20—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
    • 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)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Epidemiology (AREA)
  • Biomedical Technology (AREA)
  • Primary Health Care (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

This application provides a kind of applied to the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems, device, and method includes: the medical record information and medical diagnosis on disease result for obtaining target object；The object properties information and/or state of an illness information of target object are obtained from the medical record information of target object；The determining unmatched disease of state of an illness information with the unmatched disease of object properties information of target object and/or with target object from the medical diagnosis on disease result of target object, as the unreasonable disease in the medical diagnosis on disease result of target object.The detection method of unreasonable medical diagnosis on disease provided by the present application can detect unreasonable disease from the medical diagnosis on disease result of target object automatically, which can be improved the confidence level of medical diagnosis on disease result, so that potential medical-risk be effectively reduced.

Description

Applied to the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems, device
Technical field
This application involves medical information detection technique field more particularly to it is a kind of applied to Clinical Decision Support Systems not Reasonable medical diagnosis on disease detection method, device.
Background technique
In recent years, as medical industry technicalization, informationization, intelligence degree step up, clinical decision system (Clinical Decision Support System, CDSS) has been more and more widely used.CDSS is one based on people The medical information technical application system of machine interaction, it is intended to provide clinical decision support for doctor and other hygienic practitioners, lead to It crosses the auxiliary such as data, model and completes clinical decision.CDSS changes traditional diagnostic mode, and (i.e. dependence doctors experience and inspection refer to Target diagnostic mode), it can reduce human error's probability, optimize Allocation of Medical Resources, realize that intelligentized clinical decision refers to It leads.
The complexity with logic however, medical diagnosis is gained knowledge causes CDSS that may provide some inappropriate or serious The medical diagnosis on disease result of mistake.Some unreasonable medicals diagnosis on disease may be provided in view of CDSS as a result, need one kind can be right The disease detection result that CDSS is provided carries out the scheme of unreasonable medical diagnosis on disease detection.
Summary of the invention
In view of this, this application provides a kind of unreasonable medical diagnosis on disease detection sides applied to Clinical Decision Support Systems Method, device, to detect unreasonable medical diagnosis on disease from the disease detection result that CDSS is provided, its technical solution is as follows:
A kind of unreasonable medical diagnosis on disease detection method applied to Clinical Decision Support Systems, comprising:
Obtain the medical record information and medical diagnosis on disease result of target object；
The object properties information and/or state of an illness information of the target object are obtained from the case information；
It is determining with the unmatched disease of object properties information from the medical diagnosis on disease result, and/or with the disease The unmatched disease of feelings information, as the unreasonable disease in the medical diagnosis on disease result.
Optionally, the detection method of the unreasonable medical diagnosis on disease further include:
The unreasonable disease is deleted from the medical diagnosis on disease result；
Medical diagnosis on disease result after deleting the unreasonable disease medical diagnosis on disease result final as the target object.
Optionally, the determining and unmatched disease of object properties information from the medical diagnosis on disease result, comprising:
Group's element is obtained from the object properties information of the target object based on multiple groups predetermined, In, the multiple group is based on object property definition；
Group belonging to the target object is determined from multiple groups predetermined based on group's element, As target group；
It is determining with the unmatched disease of the target group from the medical diagnosis on disease result, as with the target object The unmatched disease of object properties information.
Optionally, the determining and unmatched disease of the target group from the medical diagnosis on disease result, comprising:
Subgraph relevant to the target group is determined from the full connexon set of graphs of group being obtained ahead of time, and is determined All subgraphs form target subgraph set；Wherein, the full connexon set of graphs of the group is based on the group relation figure constructed in advance It obtains, the group relation figure can characterize the relationship between multiple groups predetermined, the full connexon of group Any node in any subgraph in set of graphs, which represents, has connection relationship between a group, any two node；
Based on the target subgraph set, determined from the medical diagnosis on disease result unmatched with the target group Disease.
Optionally, described to be based on the target subgraph set, it is determined from the medical diagnosis on disease result and the target The unmatched disease of group, comprising:
Obtain the corresponding probability of each subgraph in the target subgraph set and disease set, wherein any subgraph is corresponding Probability by the determine the probability of each node in the subgraph, the probability of any node passes through group relevant to node element It determines；
Based on the corresponding probability of subgraph each in the target subgraph set and disease set, target disease probability set is updated It closes, wherein the target disease Making by Probability Sets includes the probability of each disease in target disease set, the target disease set Including disease involved in all case informations；
Based on updated target disease Making by Probability Sets, determined from the medical diagnosis on disease result and the target group Unmatched disease.
Optionally, described based on the corresponding probability of subgraph each in the target subgraph set and disease set, update mesh Mark disease Making by Probability Sets, comprising:
The subgraph that acquisition one had not been obtained from the target subgraph set is as target subgraph；
Based on the corresponding disease set of the target subgraph, obtained and the target subgraph phase from the group relation figure The subgraph of pass, wherein corresponding comprising the target subgraph in the corresponding disease set of relevant to target subgraph subgraph Disease in disease set；
Based on the target subgraph and the corresponding probability of subgraph relevant to the target subgraph, the target is updated The probability of disease is specified in disease Making by Probability Sets, then execution is described obtains a subgraph as mesh from the target subgraph set Subgraph is marked, until the subgraph not having not been obtained in the target subgraph set, wherein the specified disease includes at least described Disease in the corresponding disease set of target subgraph.
Optionally, it is described be based on updated target disease Making by Probability Sets, determined from the medical diagnosis on disease result with The unmatched disease of target group, comprising:
For any disease in the medical diagnosis on disease result:
The corresponding probability of the disease is obtained from the updated target disease Making by Probability Sets；
If the corresponding probability of the disease is less than predetermined probabilities threshold value, it is determined that the disease is to mismatch with the target group Disease；
With obtain in the medical diagnosis on disease result with the unmatched disease of the target group.
Optionally, the determining and unmatched disease of state of an illness information from the medical diagnosis on disease result, comprising:
When the first object disease in the medical diagnosis on disease result comprising being related to position, the disease based on the target object Feelings information determines the disease sites of the target object；
If the position that the disease sites of the target object and the first object disease are related to mismatches, it is determined that described The first object disease in medical diagnosis on disease result is and the unmatched disease of state of an illness information.
Optionally, the disease sites of the target object are determined and position that the first object disease is related to whether Match, comprising:
Disease sites and the first object disease based on the portion structure tree and the target object pre-established Position of the position being related in the portion structure tree, determine the target object disease sites and the first object disease Whether the position that disease is related to matches, wherein the portion structure tree is position hyponymy tree.
Optionally, the determining and unmatched disease of state of an illness information from the medical diagnosis on disease result, comprising:
When in the medical diagnosis on disease result including the third target disease easily obscured with the second target disease, based on described The state of an illness information of target object determines the disease that the target object is suffered from；
If the disease that the target object is suffered from is second target disease, it is determined that in the medical diagnosis on disease result The third target disease be and the unmatched disease of state of an illness information.
Optionally, the state of an illness information based on the target object determines the disease that the target object is suffered from, comprising:
In the target disease disaggregated model that the state of an illness information input is pre-established, obtain what the target object was suffered from Disease；
Wherein, it is easy to be that second target disease and the third target disease form for the target disease disaggregated model Obscure disease and classification of diseases model easily is obscured to corresponding, the target disease disaggregated model uses disease for second mesh It marks disease but is diagnosed as state of an illness sample and/or the disease of third target disease easily as the third target disease but is easily diagnosed State of an illness sample training for second target disease obtains.
A kind of unreasonable medical diagnosis on disease detection device applied to Clinical Decision Support Systems, comprising: the first acquisition module, Second obtains module and unreasonable disease detection module；
Described first obtains module, for obtaining the medical record information and medical diagnosis on disease result of target object；
Described second obtains module, for obtaining the object properties information of the target object from the case information And/or state of an illness information；
The unreasonable disease detection module, for the determining and object properties information from the medical diagnosis on disease result Unmatched disease, and/or with the unmatched disease of state of an illness information, as the unreasonable disease in the medical diagnosis on disease result Disease.
Optionally, the unreasonable disease detection module includes: first detection module；
The first detection module, for being believed based on multiple groups predetermined from the object properties of the target object Group's element is obtained in breath, wherein the multiple group is based on object property definition；
Group belonging to the target object is determined from multiple groups predetermined based on group's element, As target group；It is determining with the unmatched disease of the target group from the medical diagnosis on disease result, as with the mesh Mark the unmatched disease of object properties information of object.
Optionally, the unreasonable disease detection module includes: the second detection module；
Second detection module, for working as in the medical diagnosis on disease result comprising being related to the first object disease at position When, the disease sites of the target object are determined based on the state of an illness information of the target object；If the illness of the target object The position that position and the first object disease are related to mismatches, it is determined that the first object in the medical diagnosis on disease result Disease be and the unmatched disease of state of an illness information.
Optionally, the unreasonable disease detection module includes: third detection module；
The third detection module, in the medical diagnosis on disease result comprising easily obscure with the second target disease the When three target diseases, disease that the target object suffered from is determined based on the state of an illness information of the target object；If the target The disease that object is suffered from is second target disease, it is determined that the third target disease in the medical diagnosis on disease result is With the unmatched disease of state of an illness information.
A kind of unreasonable medical diagnosis on disease equipment applied to Clinical Decision Support Systems, comprising: memory and processor；
The memory, for storing program；
The processor realizes the unreasonable disease for being applied to Clinical Decision Support Systems for executing described program Each step of sick diagnostic method.
A kind of readable storage medium storing program for executing, is stored thereon with computer program, which is characterized in that the computer program is processed When device executes, each step of the unreasonable methods for the diagnosis of diseases applied to Clinical Decision Support Systems is realized.
Via above scheme it is found that the unreasonable medical diagnosis on disease provided by the present application applied to Clinical Decision Support Systems is examined Method, apparatus is surveyed, the object properties information and/or state of an illness information of target object can be obtained from the medical record information of target object, Then from the medical diagnosis on disease result of target object the determining unmatched disease of object properties information with target object and/or with The unmatched disease of state of an illness information of target object, to obtain the unreasonable disease in the medical diagnosis on disease result of target object. It can be seen that the unreasonable medical diagnosis on disease detection method provided by the present application applied to Clinical Decision Support Systems can be automatically from mesh Mark in the medical diagnosis on disease result of object and detect unreasonable disease, so as to obtaining the higher medical diagnosis on disease of confidence level as a result, And then reduce potential medical-risk.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis The attached drawing of offer obtains other attached drawings.
Fig. 1 is the unreasonable medical diagnosis on disease detection method provided by the embodiments of the present application applied to Clinical Decision Support Systems Flow diagram；
Fig. 2 is the determining object with target object in the medical diagnosis on disease result provided by the embodiments of the present application from target object The flow diagram of the realization process of the unmatched disease of attribute information；
Fig. 3 is that determining and target group mismatch in the medical diagnosis on disease result provided by the embodiments of the present application from target object Disease realization process flow diagram；
Fig. 4 is the Local map of the group relation figure of the multiple informative populations provided by the embodiments of the present application defined based on table 1；
Fig. 5 is an exemplary schematic diagram of group relation figure provided by the embodiments of the present application；
Fig. 6 is provided by the embodiments of the present application determining from the medical diagnosis on disease result of target object based on target subgraph set Out with the flow diagram of the realization process of the unmatched disease of target group；
Fig. 7 is an exemplary schematic diagram for portion structure tree provided by the embodiments of the present application；
Fig. 8 is the unreasonable medical diagnosis on disease detection device provided by the embodiments of the present application applied to Clinical Decision Support Systems Structural schematic diagram；
Fig. 9 is the unreasonable medical diagnosis on disease detection device provided by the embodiments of the present application applied to Clinical Decision Support Systems Structural schematic diagram.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete Site preparation description, it is clear that described embodiments are only a part of embodiments of the present application, instead of all the embodiments.It is based on Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other Embodiment shall fall in the protection scope of this application.
Medical diagnosis is gained knowledge and logical comparison complexity, is embodied in: accurate medical diagnosis on disease is in addition to being related to patient's Symptom, checks the medical features such as test rating, also related to many factors such as the inducement of morbidity, group, working environment；No May have a similar symptom and performance with disease, acute a variety of situations such as chronic of disease are also required to take different disposition sides Formula；Part emergency treatment, surgical diagnosis are also related to the specific injury position of patient, injury mode and injury degree.Above-mentioned diagnosis is multiple Polygamy may result in CDSS and provide unreasonable medical diagnosis on disease.
CDSS provides unreasonable medical diagnosis on disease, on the one hand can reduce medical personnel for the degree of belief of CDSS, in turn Cause the retroversion of medical diagnostic modalities, on the other hand, unreasonable medical diagnosis on disease may result in patient is taken it is inappropriate Therapeutic modality is delayed best occasion for the treatment, causes serious consequence.In order to avoid the generation of above situation, it is necessary to CDSS The diagnostic result provided is checked that current test mode is mostly artificial sampling observation mode, specifically, sampling observation personnel pass through first The sampling observation rule of setting, retrieves the case history and corresponding diagnostic message of patient, then relies on the judgement mark of personal experience and setting Standard checks the integrity degree of element, the compliance of content, diagnosis reasonability etc. in case history.
Inventor it has been investigated that: the sampling observation effect for the mode manually inspected by random samples depend on hospital rules and regulations and pumping The personal experience of inspection personnel and subjective initiative however, different grades of hospital sampling observation Policy Difference is larger, also, inspect people by random samples Member's quality is very different, this influences the detection quality of unreasonable medical diagnosis on disease very big.In addition, the mode using sampling observation carries out not The detection of reasonable medical diagnosis on disease can not cover all cases, this brings security risk to hospital and patient.
In view of existing artificial sampling observation mode there are the above problem, inventor is had made intensive studies, final to propose The detection method of the preferably unreasonable medical diagnosis on disease of detection effect a kind of, this method can automatically examine the disease of CDSS output Disconnected result carries out the detection of unreasonable medical diagnosis on disease, which is applicable to need the medical diagnosis on disease result to diseased subjects The application scenarios of unreasonable medical diagnosis on disease detection are carried out, which can be applied to terminal, can also be applied to server, connect down To be introduced by detection method of following embodiments to unreasonable medical diagnosis on disease provided by the present application.
Referring to Fig. 1, showing the unreasonable disease provided by the embodiments of the present application applied to Clinical Decision Support Systems The flow diagram of diagnosis detecting method, this method may include:
Step S101: the medical record information and medical diagnosis on disease result of target object are obtained.
Wherein, the case information of target object is about the generation of target object illnesses, development, diagnosis, treatment feelings The record of condition, the medical diagnosis on disease result of target object can be the medical diagnosis on disease that CDSS is provided based on the state of an illness information of target object As a result.
Step S102: the object properties information and/or state of an illness letter of target object are obtained from the case information of target object Breath.
Including at least the object properties information of target object and the state of an illness of target object letter in the case information of target object Breath.Wherein, the object properties information of target object can be the personal essential information of target object, for example, gender, age, duty Industry, height, weight etc., the state of an illness information of target object are the relevant information of target object illnesses.
Step S103: the determining object properties information with target object mismatches from the medical diagnosis on disease result of target object Disease, and/or the unmatched disease of state of an illness information with target object, as in the medical diagnosis on disease result of target object not Reasonable disease.
Optionally, the present embodiment can pass through the object category of disease and target object in the medical diagnosis on disease result of target object Whether property information matches to determine unreasonable disease.Illustratively, it is male that the object properties information of target object, which includes: gender, Property, the age be 65 one full year of life, it is assumed that have pediatric disease or gynecological disease in the medical diagnosis on disease result of target object, it is clear that paediatrics disease Disease and the age of target object mismatch, and the gender of gynecological disease and target object mismatches, then, it may be determined that medical diagnosis on disease knot Pediatric disease or gynecological disease in fruit are unreasonable disease.
Optionally, the present embodiment can also pass through the state of an illness of disease and target object in the medical diagnosis on disease result of target object Whether information matches to determine unreasonable medical diagnosis on disease.Illustratively, the corresponding disease of state of an illness information of target object is A, and Disease in the medical diagnosis on disease result of target object is but B, it is clear that disease B is unreasonable medical diagnosis on disease, alternatively, target object State of an illness information in the disease sites of target object be C, and the position that the disease in the medical diagnosis on disease result of target object is related to For D, then the disease for being related to position D is unreasonable medical diagnosis on disease.
Preferably, the present embodiment can be determined from the medical diagnosis on disease result of target object believes with the object properties of target object Cease unmatched disease and the unmatched disease of state of an illness information with target object, the medical diagnosis on disease result as target object In unreasonable disease.
Unreasonable medical diagnosis on disease detection method provided by the embodiments of the present application applied to Clinical Decision Support Systems, can be from The object properties information and/or state of an illness information that target object is obtained in the medical record information of target object, then from target object It is determining in medical diagnosis on disease result to believe with the unmatched disease of object properties information of target object and/or with the state of an illness of target object Unmatched disease is ceased, to obtain the unreasonable disease in the medical diagnosis on disease result of target object.It can be seen that the application is real The unreasonable medical diagnosis on disease detection method applied to Clinical Decision Support Systems for applying example offer can be automatically from the disease of target object Unreasonable disease is detected in sick diagnostic result, so as to obtain the higher medical diagnosis on disease of confidence level as a result, reducing in turn latent Medical-risk, in addition, the detection method avoids hospital's sampling observation specification and sampling observation personnel compared to artificial sampling observation mode The influence of subjective factor, and all cases can be covered comprehensively.
In another embodiment of the application, to " the step S103: from the medical diagnosis on disease of target object in above-described embodiment As a result the object properties information and/or the unmatched disease of state of an illness information of middle determination and target object, are examined as unreasonable disease It is disconnected " realization process be introduced.
First to determining unmatched with the object properties information of target object from the medical diagnosis on disease result of target object The realization process of disease is introduced, referring to Fig. 2, showing the flow diagram of the realization process, may include:
Step S201: group is obtained from the object properties information of target object based on multiple groups predetermined and is wanted Element.
In the present embodiment, the multiple groups of object property definition can be in advance based on, the following table 1 shows predetermined multiple One example of group:
The multiple groups predetermined of table 1
It should be noted that several not mutual exclusions can be belonged to according to its own situation for any object Group, for example, an object can belong to " 2 " (25~45 years old age) of table, " 3 " (women), " 15 " (occupation: light), " 20 " (155~165cm of height) etc., in addition, same target can belong to two similar populations, for example, can belong to " 39 " (newborn) " 41 " (child) can be belonged to.
In the present embodiment, target pair can be extracted from the object properties information of target object based on group predetermined Age of elephant, gender, pregnant no, professional, native place, permanent residence, by information such as ground as group's element.Optionally, base can be used Group's element is extracted from the object properties information of target object in the Entity recognition mode of keyword, rule, NER etc..It needs Illustrate, corresponding for each group's element of extraction to obtain a probability, the corresponding probability of any group's element is used for table Levy group's element establishment confidence level, for example, patient make a definite diagnosis pregnancy 3 weeks, then identify patient be pregnant woman probability be 0.8, Pregnant woman is exactly group's element, and the confidence level that this group's element is set up is 0.8.
Step S202: group belonging to target object is determined from multiple groups predetermined based on group's element.
Specifically, the corresponding group of group's element that step S201 is obtained is determined from multiple groups predetermined, it will The corresponding group of group's element is determined as group belonging to target object.
Illustratively, group's element of target object are as follows: 28 years old age, gender women, pregnant woman, height 160, permanent residence are green Island would know that 28 years old corresponding group is " 2 " (25~45 years old age) by upper table 1, and the corresponding group of women is " 3 " (female Property), the corresponding group of pregnant woman is " 5 " (pregnancy period), and the corresponding group of height 160 is " 20 " (155~165cm of height) and " 22 " (160~170cm), the corresponding group in permanent residence Qingdao are " 33 " (life is regional: city) and " 37 " (life area: coastal), That is, group belonging to target object include " 2 " (25~45 years old age), " 3 " (women), " 5 " (pregnancy period), " 20 " (height 155~ 165cm), " 22 " (160~170cm), " 33 " (life area: city), " 37 " (life area: coastal).
Step S203: it is determining with the unmatched disease of target group from the medical diagnosis on disease result of target object, as with The unmatched disease of object properties information of target object.
Illustratively, group belonging to target object includes " 2 " (25~45 years old age) in table 1, the disease of target object It include pediatric disease in sick diagnostic result, it is clear that the age of pediatric disease and target object mismatches, therefore, target object Pediatric disease in medical diagnosis on disease result is and group " 2 " (25~45 years old age) unmatched disease.
Further below in above-mentioned steps S203 from the medical diagnosis on disease result of target object it is determining with target group not The realization process of matched disease is introduced, referring to Fig. 3, showing the flow diagram of the realization process, may include:
Step S301: subgraph relevant to target group is determined from the full connexon set of graphs of group being obtained ahead of time, really All subgraphs composition target subgraph set made.
Wherein, the full connexon set of graphs of group is obtained based on the group relation figure constructed in advance, and group relation figure is characterization The figure of relationship between multiple groups predetermined.Referring to Fig. 4, showing the group of the multiple informative populations defined based on table 1 The Local map of body relational graph, group relation figure is based on whether having direct relation building between any two group, for example, " male " It is not directly dependent upon with " women ", then in building group's relational graph, nothing is directly connected between " male " and " women ", then is compared Such as, male and female are likely to allergy, therefore, are directly connected between " male " and " allergy ", between " women " and " allergy " Also it is directly connected to, other groups are similar.
After obtaining group relation figure, the full connexon set of graphs of group can be obtained based on group relation figure, it is complete for group Any subgraph in connexon set of graphs, any node in the subgraph, which represents to have between a group, any two node, to be connected Connect relationship.
Assumed group relational graph is G, and the node set in group relation figure is E, and the line set in group relation figure is V, Then the process of the acquisition full connexon set of graphs G ' of group may include: from group relation figure G
Step 1) traverses line set V: for the side currently traversed, being added the node at its both ends as a node group Full connexon set of graphs G ' (initial full connexon set of graphs G ' is null set), then traverses lower a line, until having traversed V In all sides；
Step 2) traverse node set E: for the node currently traversed, from the middle determination of current G ' with currently traverse Node meets the node group of preset condition, for each node group for meeting preset condition, by the node group node with work as Before the node that traverses form a node group G ' be added, next node is then traversed, until traverse owning in node set E Then node executes step 2), until not having new node group to be added in G '.It should be noted that in the G ' finally obtained Any node group represents a full connection subgraph.
Wherein, the node group of preset condition is met with the node currently traversed are as follows: exist with the node currently traversed The node group on side, it should be noted that there are sides to refer in the node group for a node group and the node currently traversed Each node with there are sides between the node that currently traverses.
Referring to Fig. 5, showing the example of group's relational graph, by traversing the line set of group's relational graph, can get The full connexon set of graphs of group { (pregnancy period, women), (women, child), (women, 5~15 years old age), (male, child), (male Property, 5~15 years old age), (child, 5~15 years old age) }, further on the basis of above-mentioned set, traversal group's relational graph Node set can get new node group (women, child, 5~15 years old age) and (male, child, 5~15 years old age), then most The full connexon atlas of the group obtained eventually is combined into { (pregnancy period, women), (women, child), (women, 5~15 years old age), (male Property, child), (male, 5~15 years old age), (child, 5~15 years old age), (women, child, 5~15 years old age), (male, Child, 5~15 years old age) }.
It should be noted that the mode that above-mentioned acquisition group connects subgraph entirely is only an example, the present embodiment is not limited Through the above way obtain group connect subgraph entirely, group connects subgraph entirely and can also be achieved other ways, for example, can first by Full connexon set of graphs G ' is added as a node group in the node at each edge both ends in group relation figure, then, closes from group It is that the side that can make up closed polygon is searched in figure, G ' is added in the node group that will form the side of closed polygon, for example, In Fig. 5, the side between women and child, women and age~5~15 year old, child and age~5~15 year old can form closure three It is angular, then G ' is added in the node group of women, child and age~5~15 year old composition.
After obtaining the full connexon set of graphs of group, determination is relevant to target group from group's full connexon set of graphs Subgraph, it should be noted that any subgraph relevant to target group meets: all nodes in the subgraph are target group In group.Illustratively, the full connexon atlas of group be combined into (pregnancy period, women), (women, child), (women, the age 5~ 15 years old), (male, child), (male, 5~15 years old age), (child, 5~15 years old age), (women, child, age 5~15 Year), (male, child, 5~15 years old age), it is assumed that target group include male, child, the age 5~15 years old, then with target complex The relevant subgraph of body includes (male, child), (male, 5~15 years old age), (male, child, 5~15 years old age).
Step S302: it is determined with target group not from the medical diagnosis on disease result of target object based on target subgraph set Matched disease.
Specifically, being determined from the medical diagnosis on disease result of target object referring to Fig. 6, showing based on target subgraph set Out with the flow diagram of the realization process of the unmatched disease of target group, may include:
Step S601: the corresponding probability of each subgraph and disease set in target subgraph set are obtained.
Wherein, the corresponding disease set of any subgraph can be predefined based on expertise, the corresponding probability of any subgraph By the determine the probability of each node in the subgraph, the probability of any node is determined by group relevant to node element, It is used to characterize the confidence level that relevant group's element is set up.
In one possible implementation, the corresponding probability of any subgraph can be the probability of each node in the subgraph In minimum probability, that is, the probability m (g ') of any subgraph g ' are as follows:
Wherein, z is normaliztion constant, and G ' is the full connexon set of graphs of group, and G'* is that (i.e. group is complete for target subgraph set In connexon set of graphs, the set of subgraph composition relevant to group belonging to target object), g ' is in target subgraph set One subgraph, pi1,pi2,pi3... for the probability of each node in g '.
Step S602: based on each corresponding probability of subgraph and the corresponding disease collection of each subgraph in target subgraph set It closes, updates target disease Making by Probability Sets.
Wherein, target disease Making by Probability Sets includes the corresponding probability of each disease, target disease collection in target disease set I The disease closed in I can be for disease involved in all case informations, i.e. all diseases that CDSS is supported.
The corresponding probability P of any disease i in target disease set I0(i) it can be determined by following formula:
Wherein, count (*) identifies the frequency that * occurs in all cases.
Specifically, based on each corresponding probability of subgraph and the corresponding disease collection of each subgraph in target subgraph set It closes, the process for updating target disease Making by Probability Sets may include:
Step S6021, the subgraph that acquisition one had not been obtained from target subgraph set is as target subgraph.
Step S6022, it is based on the corresponding disease set of target subgraph, is obtained from group relation figure related to target subgraph Subgraph.
Specifically, target subgraph pair, the non-targeted son for the target subgraph centering that will acquire can be obtained from group relation figure Figure is as subgraph relevant to target subgraph, wherein the corresponding disease intersection of sets of two subgraphs of target subgraph centering Integrate as the corresponding disease set of target subgraph.
Illustratively, the corresponding disease collection of target subgraph Z is combined into { i1, i2, it is assumed that in group relation figure there are subgraph X and The corresponding disease collection of subgraph Y, subgraph X is combined into { i1, i2, i3, the corresponding disease collection of subgraph Y is combined into { i1, i2, i4, due to subgraph X Corresponding disease set { i1, i2, i3Disease set { i corresponding with subgraph Y1, i2, i4Intersection be { i1, i2, therefore, subgraph X and subgraph Y is subgraph relevant to target subgraph.
Step S6023, target subgraph and the corresponding probability of subgraph relevant to target subgraph, more fresh target disease are based on The probability of disease is specified in sick Making by Probability Sets.
Wherein, disease is specified to include at least the disease in the corresponding disease set of target subgraph.
Specifically, using target subgraph and the corresponding probability of subgraph relevant to target subgraph, more using following formula The probability of target disease in fresh target disease Making by Probability Sets:
Wherein, B' and C' is any two subgraph in group relation figure G, B' and the corresponding disease intersection of sets of C' Integrate as the corresponding disease set I ' of subgraph g ', p (d) is that any disease d in the corresponding disease set I ' of target subgraph g' is corresponding Probability.
Step S6024, the subgraph whether also having not been obtained in target subgraph set is judged, if so, thening follow the steps S6021, if it is not, then terminating to update.
Below by a specific example to based on each corresponding probability of subgraph and each subgraph in target subgraph set Corresponding disease set, the process for updating target disease Making by Probability Sets are illustrated:
Assumed group relational graph is G, target subgraph set G'*={ g1, g2, g3, g4, g1Corresponding disease set I1= {i1, g2Corresponding disease set I2={ i1, i2, g3Corresponding disease set I3={ i3, g4Corresponding disease set I4= {i3, i4, g1Corresponding probability is m (g1)=p1, g2Corresponding probability is m (g2)=p2, g2Corresponding probability m (g3)=p3, g4 Corresponding probability is m (g4)=p4:
(1) it is based on g1Carry out first time update:
Due to I2With I1Intersection be I1, then it is based on g1And g2Corresponding probability and disease set update target disease probability Set, specific:
m1(g1)=p1, m1(g2)=p2, m1(G)=1-p1-p2, m1(other subgraphs)=0 initializes M0(G)=1, M0It (removes Other outer subgraphs of G)=0, then:
K=M0(g1)*M0(g3)+M0(g1)*M0(g4)+M0(g2)*M0(g3)+M0(g2)*M0(g4)=0
M1(g3)=0M1(g4)=0
M1(G)=1-m1(g1)-m1(g2)
Then, i is updated1、i2、i3、i4Corresponding probability:
(2) it is based on g3Second is carried out to update:
Due to I4With I3Intersection be I3, then it is based on g3And g4Corresponding probability and disease set update target disease probability Set, specific:
m2(g3)=p3, m2(g4)=p4, m2(G)=1-p3-p4, m2(other subgraphs)=0, then:
K=M1(g1)*M1(g3)+M1(g1)*M1(g4)+M1(g2)*M1(g3)+M1(g2)*M1(g4)=0
M2(G)=1- (1-m1(g1)-m1(g2))*m2(g4)-(1-m1(g1)-m1(g2))*m2(g3)
Then, i is updated1、i2、i3、i4Corresponding probability:
Step S603: being based on updated target disease Making by Probability Sets, determines from the medical diagnosis on disease result of target object Out with the unmatched disease of target group.
Specifically, for any disease in the medical diagnosis on disease result of target object: from updated target disease probability In set, the corresponding probability of the disease is obtained；If the corresponding probability of the disease is less than predetermined probabilities threshold θ, it is determined that the disease For with the unmatched disease of target group, i.e. the disease is unreasonable medical diagnosis on disease, and the disease of so available target object is examined Unreasonable disease in disconnected result.
It is carried out below to determining from the diagnostic result of target object with the target object unmatched disease of state of an illness information It introduces.
It is related to the medical diagnosis on disease at position in view of that may exist in the medical diagnosis on disease result of target object, for example, hand bone Folding, upper limb damage etc., the medical diagnosis on disease for being related to position may have the position that medical diagnosis on disease is related to and diseased subjects The unmatched situation in actual diseased position, in view of this, in one possible implementation, from the diagnostic result of target object The process of the unmatched disease of the state of an illness information of middle determination and target object may include:
When the first object disease in the medical diagnosis on disease result of target object comprising being related to position, based on target object State of an illness information determines the disease sites of target object；If the position that the disease sites of target object and first object disease are related to is not Matching, it is determined that first object disease is the unmatched disease of state of an illness information with target object, i.e., first object disease is not Reasonable medical diagnosis on disease.
Wherein it is determined that the whether matched process in position that is related to of the disease sites of target object and first object disease can be with It include: the position that disease sites and first object disease based on the portion structure tree and target object pre-established are related to Position in portion structure tree, determines whether the disease sites of target object match with the position that first object disease is related to.
It should be noted that portion structure tree is human body hyponymy tree, portion structure tree includes multiple positions Tree node, each position tree node represent a position, and multiple position tree nodes have hyponymy, it is preferred that position knot Paper mulberry can also include the synonym tree node of position tree node, referring to Fig. 7, showing an example for portion structure tree, Fig. 7 The portion structure tree shown includes the position burl for respectively representing palm, finger, hand, arm, upper limb, lower limb, four limbs and trunk Point further includes the synonym tree node for representing arm, and as seen from Figure 7, palm and finger are the bottom of hand, and hand and arm are The bottom of upper limb, upper limb and lower limb are the bottom of four limbs, and four limbs are the bottom of trunk, and arm is the synonym of arm.
Further, disease sites and first object disease based on the portion structure tree and target object pre-established Position of the position that disease is related in portion structure tree determines the portion that the disease sites of target object and first object disease are related to Position whether matched process include: the determination target object from portion structure tree the corresponding tree node of disease sites, as the One tree node；The corresponding tree node in position that first object disease is related to is determined from portion structure tree, as the second tree node； First tree node is matched with the second tree node, if the first tree node is identical as the second tree node or the first tree node For the synonym tree node of the second tree node, alternatively, the second tree node is the upper tree node of the first tree node, then the can determine The position that one target disease is related to is matched with the disease sites of target object；Otherwise, it determines the position that first object disease is related to It is mismatched with the disease sites of target object, i.e., first object disease is unreasonable medical diagnosis on disease.Illustratively, target object Disease sites are palm, and medical diagnosis on disease result is upper limb damage, since upper limb is the upper of palm, upper limb damage is conjunction The medical diagnosis on disease of reason.
It wherein, can be first by illness in the corresponding tree node of the disease sites of target object determining from portion structure tree The title at position is normalized to title, for example, disease sites are arm, can be normalized to arm, is obtaining illness After the title at position, the corresponding tree node of disease sites is determined from portion structure tree based on the title.Optionally, When being matched the first tree node with the second tree node, the first tree node and the second tree node can be subjected to recurrence matching, Specifically, portion structure tree, bottom-up matching can be based on.
It is understood that can have the same or similar disease of some symptoms in numerous diseases, for example, disease A with The symptom of disease B is similar, then is easy to for disease A to be diagnosed as disease B in diagnosis, or disease B is diagnosed as disease A, have In consideration of it, in one possible implementation, the determining state of an illness information with target object from the diagnostic result of target object The process of unmatched disease may include:
When in the medical diagnosis on disease result of target object including the third target disease easily obscured with the second target disease, base The disease that target object is suffered from is determined in the state of an illness information of target object；If the disease that target object is suffered from is the second target disease Disease, it is determined that the third target disease in the medical diagnosis on disease result of target object is unmatched with the state of an illness information of target object Disease, i.e. third target disease in the medical diagnosis on disease result of target object are unreasonable disease.
Illustratively, disease A is easily to obscure disease, is easy to be confused with disease B, it is assumed that pass through the state of an illness of target object Information determines that the disease of target object is disease A, and the medical diagnosis on disease result of target object is disease B, then target object Medical diagnosis on disease result in disease B be unreasonable disease.
It is understood that easily obscure whether disease is unreasonable medical diagnosis on disease to determine in medical diagnosis on disease result, Firstly the need of determining whether containing easily disease is obscured in medical diagnosis on disease result, in the embodiment of the present application, can be obtained ahead of time easily mixed The disease pair confused is then based on the confusing disease that is obtained ahead of time to whether containing in the medical diagnosis on disease result for determining target object Have and easily obscures disease.
Easily obscure the process of disease pair below and be introduced to being obtained ahead of time:
Assuming that the disease that CDSS is supported is N kind, the state of an illness sample set S of N kind disease is obtained, the sample number of every kind of disease is K then shares N*K sample in state of an illness sample set S, it is the matrix of N*N that disease confusion matrix, which is M, the either element in M MijAre as follows:
Wherein, KijDisease for sample is i but is diagnosed as the quantity of the sample of disease j, likewise, KjiFor sample Disease be j's but be diagnosed as disease i sample quantity.
If MijGreater than the threshold gamma of obscuring of setting, then (disease i, disease j) are easily to obscure disease pair, so available N kind Disease pair is easily obscured in disease.
Disease easily is obscured to rear obtaining, and easily obscures disease to (disease i, disease j), can be used K for anyijUKjiInstruction Practice two classifiers, obtains and easily obscure disease to (the corresponding classification of diseases model of disease i, disease j), in this way, can get each easy Disease is obscured to corresponding classification of diseases model.
Medical diagnosis on disease for target object is as a result, when it includes the third target diseases easily obscured with the second target disease When, based on target object state of an illness information determine target object disease process may include: can be by the state of an illness of target object Information input disease obtains target object and is suffered to (the second target disease, third target disease) corresponding classification of diseases model Disease (i.e. the second target disease or third target disease).Wherein, (the second target disease, third target disease) corresponding disease Sick disaggregated model uses disease for the second target disease but is easily diagnosed as the state of an illness sample of third target disease and/or disease is It third target disease but is easily diagnosed as the state of an illness sample training of the second target disease and obtains.
Above scheme gives the realization process detected from medical diagnosis on disease result of three angles to target object, and three Whether the disease that a angle is respectively as follows: in the medical diagnosis on disease result of (1) target object matches with group belonging to target object； (2) whether the position that the disease in the medical diagnosis on disease result of target object is related to matches with the disease sites of target object；(3) mesh Whether the disease marked in the medical diagnosis on disease result of object is the disease being confused.
For each disease in the medical diagnosis on disease result of target object, the embodiment of the present application is preferably from above three angle Degree detects the disease, so that three testing results are obtained for the disease, then to three testing results of the disease It is merged, final detection result of the fused result as the disease.
For each disease in the medical diagnosis on disease result of target object, it is assumed that three testing results of the disease are respectively d1、d2、d3, then the final detection result D of the disease are as follows:
D=fun (d1,d2,d3) (6)
Wherein, fun is fusion treatment function.In one possible implementation, fusion treatment function can be any Boolean calculation function.
It is illustrative:
fun(d1,d2,d3)=d1and d2and d3 (7)
For any disease in the medical diagnosis on disease result of target object:
If group belonging to the disease and target object mismatches, it is determined that the first testing result of the disease is 0, with 0 Characterizing the disease is unreasonable disease, conversely, the first testing result for defining the disease is 1, characterizing the disease with 1 is reasonable disease Disease；If the disease is not to be related to the disease at position, the second testing result for defining the disease is 1, if the disease is to be related to portion The disease of position, and the disease sites at position and target object that the disease is related to mismatch, then define the second detection of the disease It as a result is 0, if the disease is to be related to the disease at position, and the position that the disease is related to is matched with the disease sites of target object, The second testing result for then defining the disease is 1；If the disease is not easily to obscure disease, the third detection knot of the disease is defined Fruit is 1, if the disease is easily to obscure disease, and the disease that the disease and target object are suffered from is inconsistent, then defines the disease Third testing result is 0, if the disease is easily to obscure disease, and the disease is consistent with target object illnesses, then definition should The third testing result of disease is 1.Obtain the disease three testing results after, can be used above formula to three testing results into Row fusion, if fun (d1,d2,d3)=0, it is determined that the disease is unreasonable disease, if fun (d1,d2,d3)=1, it is determined that should Disease is reasonable disease.
That is, for any disease in the medical diagnosis on disease result of target object, if the disease meets the following conditions In any one or any two, or meet three simultaneously, it is determined that the disease is unreasonable medical diagnosis on disease: (1) should Group belonging to disease and target object mismatches；(2) disease sites at the position and target object that the disease is related to mismatch； (3) disease, which is confused with, easily obscures disease, and its disease determining with the state of an illness information based on target object is inconsistent.
It can get the unreasonable medical diagnosis on disease in the medical diagnosis on disease result of target object via the above process, it is basic herein On, unreasonable methods for the diagnosis of diseases provided by the embodiments of the present application can also include: not conform to what is determined via the above process Reason disease is deleted from the medical diagnosis on disease result of target object, and the medical diagnosis on disease result after deleting unreasonable disease is as target pair As final medical diagnosis on disease result.
Unreasonable medical diagnosis on disease detection method provided by the embodiments of the present application applied to Clinical Decision Support Systems can be certainly It is dynamic to detect unreasonable disease from the medical diagnosis on disease result of target object, so as to obtain the higher medical diagnosis on disease of confidence level As a result, potential medical-risk is effectively reduced in turn, in addition, the detection method avoids hospital compared to artificial sampling observation mode The influence of sampling observation specification and sampling observation personnel's subjective factor, and all cases can be covered comprehensively.
The embodiment of the present application also provides a kind of unreasonable medicals diagnosis on disease applied to Clinical Decision Support Systems to detect dress It sets, the unreasonable medical diagnosis on disease detection device provided by the embodiments of the present application applied to Clinical Decision Support Systems is carried out below Description, the unreasonable medical diagnosis on disease detection device described below applied to Clinical Decision Support Systems and above-described application Reference can be corresponded to each other in the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems.
Referring to Fig. 8, showing provided by the embodiments of the present application a kind of applied to the unreasonable of Clinical Decision Support Systems The structural schematic diagram of medical diagnosis on disease detection device, the apparatus may include: first, which obtains module 801, second, obtains 802 and of module Unreasonable disease detection module 803.
First obtains module 801, for obtaining the medical record information and medical diagnosis on disease result of target object.
Second obtain module 802, for obtained from the case information target object object properties information and/ Or state of an illness information.
Unreasonable disease detection module 803, for the determining and object properties information from the medical diagnosis on disease result Unmatched disease, and/or with the unmatched disease of state of an illness information, as the unreasonable disease in the medical diagnosis on disease result Disease.
Unreasonable medical diagnosis on disease detection device provided by the embodiments of the present application applied to Clinical Decision Support Systems, can be from The object properties information and/or state of an illness information that target object is obtained in the medical record information of target object, then from target object It is determining in medical diagnosis on disease result to believe with the unmatched disease of object properties information of target object and/or with the state of an illness of target object Unmatched disease is ceased, to obtain the unreasonable medical diagnosis on disease in the medical diagnosis on disease result of target object.It can be seen that this Shen Please embodiment provide the unreasonable medical diagnosis on disease detection device applied to Clinical Decision Support Systems can be automatically from target object Medical diagnosis on disease result in detect unreasonable disease, so as to obtain the higher medical diagnosis on disease of confidence level as a result, dropping in turn Low medical-risk, in addition, the detection method avoids hospital's sampling observation specification and sampling observation personnel master compared to artificial sampling observation mode The influence of sight factor, and all cases can be covered comprehensively.
Optionally, the unreasonable medical diagnosis on disease detection device provided by the above embodiment applied to Clinical Decision Support Systems It can also include: unreasonable medical diagnosis on disease removing module.
Unreasonable medical diagnosis on disease removing module, for deleting the unreasonable disease from the medical diagnosis on disease result； Medical diagnosis on disease result after deleting the unreasonable disease medical diagnosis on disease result final as the target object.
In one possible implementation, provided by the above embodiment applied to the unreasonable of Clinical Decision Support Systems Unreasonable disease detection module 803 in medical diagnosis on disease detection device may include: first detection module.
First detection module, for based on multiple groups predetermined from the object properties information of the target object Obtain group's element, wherein the multiple group is based on object property definition；
Group belonging to the target object is determined from multiple groups predetermined based on group's element, As target group；It is determining with the unmatched disease of the target group from the medical diagnosis on disease result, as with the mesh Mark the unmatched disease of object properties information of object.
In one possible implementation, first detection module may include: for from the medical diagnosis on disease result The determining and unmatched mismatch disease determining module of the target group.
Mismatch disease determining module, specifically for from the full connexon set of graphs of group being obtained ahead of time determine with it is described The relevant subgraph of target group, all subgraphs composition target subgraph set determined；Based on the target subgraph set, from institute It states and is determined in medical diagnosis on disease result and the unmatched disease of the target group.
Wherein, the full connexon set of graphs of the group is obtained based on the group relation figure constructed in advance, the group relation Figure can characterize the relationship between multiple groups predetermined, any subgraph in the full connexon set of graphs of group In any node represent between a group, any two node have connection relationship.
In one possible implementation, mismatching disease determining module may include: acquisition submodule, probability updating Submodule and mismatch disease determine submodule.
Acquisition submodule, for obtaining the corresponding probability of each subgraph in the target subgraph set and disease set, In, the corresponding probability of any subgraph by the determine the probability of each node in the subgraph, the probability of any node by with the section The relevant group's element of point determines；
Probability updating submodule, for based on the corresponding probability of subgraph each in the target subgraph set and disease collection It closes, updates target disease Making by Probability Sets, wherein the target disease Making by Probability Sets includes each disease in target disease set Probability, the target disease set include disease involved in all case informations.
It mismatches disease and determines submodule, for being based on updated target disease Making by Probability Sets, from the medical diagnosis on disease As a result it is determined in and the unmatched disease of the target group.
In one possible implementation, probability updating submodule, specifically for being obtained from the target subgraph set Take a subgraph having not been obtained as target subgraph；Based on the corresponding disease set of the target subgraph, from the group relation Relevant to target subgraph subgraph is obtained in figure, wherein the corresponding disease set of relevant to target subgraph subgraph In include the disease in the corresponding disease set of the target subgraph；Based on the target subgraph and related to the target subgraph The corresponding probability of subgraph, update the probability that disease is specified in the target disease Making by Probability Sets, then execute it is described from A subgraph is obtained in the target subgraph set as target subgraph, until do not had not been obtained in the target subgraph set Subgraph, wherein the specified disease includes at least the disease in the corresponding disease set of the target subgraph.
In one possible implementation, probability updating submodule is being based on the subgraph and subgraph relevant to the subgraph Corresponding probability is specifically used for when updating the probability for specifying disease in the target disease Making by Probability Sets for the disease Any disease in sick diagnostic result: the corresponding probability of the disease is obtained from the updated target disease Making by Probability Sets； If the corresponding probability of the disease be less than predetermined probabilities threshold value, it is determined that the disease be and the unmatched disease of the target group； With obtain in the medical diagnosis on disease result with the unmatched disease of the target group.
In one possible implementation, in the detection device of unreasonable medical diagnosis on disease provided by the above embodiment not Reasonable disease detection module 803 may include: the second detection module.
Second detection module, for working as in the medical diagnosis on disease result comprising being related to the first object disease at position When, the disease sites of the target object are determined based on the state of an illness information of the target object；If the illness of the target object The position that position and the first object disease are related to mismatches, it is determined that the first object in the medical diagnosis on disease result Disease be and the unmatched disease of state of an illness information.
In one possible implementation, unreasonable disease detection module 803, is also used to based on the position pre-established The position that the disease sites and the first object disease of structure tree and the target object are related to is in the portion structure tree In position, determine whether the disease sites of the target object match with the position that the first object disease is related to, wherein The portion structure tree is human body hyponymy tree.
In one possible implementation, provided by the above embodiment applied to the unreasonable of Clinical Decision Support Systems Unreasonable disease detection module 803 in medical diagnosis on disease detection device may include: third detection module.
The third detection module, in the medical diagnosis on disease result comprising easily obscure with the second target disease the When three target diseases, disease that the target object suffered from is determined based on the state of an illness information of the target object；If the target The disease that object is suffered from is second target disease, it is determined that the third target disease in the medical diagnosis on disease result is With the unmatched disease of state of an illness information.
In one possible implementation, third detection module determines institute in the state of an illness information based on the target object When stating the disease that target object is suffered from, specifically for the target disease disaggregated model for pre-establishing the state of an illness information input In, obtain the disease that the target object is suffered from.
Wherein, the object-class model is that second target disease and the easy of third target disease composition are obscured Disease easily obscures classification of diseases model to corresponding, and the target disease disaggregated model uses disease for the second target disease It disease but is easily diagnosed as state of an illness sample and/or the disease of third target disease as the third target disease but is easily diagnosed as institute The state of an illness sample training for stating the second target disease obtains.
The embodiment of the present application also provides a kind of unreasonable medical diagnosis on disease detections applied to Clinical Decision Support Systems to set Standby, referring to Fig. 9, showing the structural schematic diagram of the equipment, which may include: at least one processor 901, at least one A communication interface 902, at least one processor 903 and at least one communication bus 904；
In the embodiment of the present application, processor 901, communication interface 902, memory 903, communication bus 904 quantity be At least one, and processor 901, communication interface 902, memory 903 complete mutual communication by communication bus 904；
Processor 901 may be a central processor CPU or specific integrated circuit ASIC
(Application Specific Integrated Circuit), or be arranged to implement of the invention real Apply one or more integrated circuits etc. of example；
Memory 903 may include high speed RAM memory, it is also possible to further include nonvolatile memory (non- Volatile memory) etc., a for example, at least magnetic disk storage；
Wherein, memory is stored with program, the program that processor can call memory to store, and described program is used for:
Obtain the medical record information and medical diagnosis on disease result of target object；
The object properties information and/or state of an illness information of the target object are obtained from the case information；
It is determining with the unmatched disease of object properties information from the medical diagnosis on disease result, and/or with the disease The unmatched disease of feelings information, as the unreasonable disease in the medical diagnosis on disease result.
Optionally, the refinement function of described program and extension function can refer to above description.
The embodiment of the present application also provides a kind of readable storage medium storing program for executing, which can be stored with and hold suitable for processor Capable program, described program are used for:
Obtain the medical record information and medical diagnosis on disease result of target object；
The object properties information and/or state of an illness information of the target object are obtained from the case information；
It is determining with the unmatched disease of object properties information from the medical diagnosis on disease result, and/or with the disease The unmatched disease of feelings information, as the unreasonable disease in the medical diagnosis on disease result.
Optionally, the refinement function of described program and extension function can refer to above description.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged Except there is also other identical elements in the process, method, article or apparatus that includes the element.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other The difference of embodiment, the same or similar parts in each embodiment may refer to each other.
The foregoing description of the disclosed embodiments makes professional and technical personnel in the field can be realized or use the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein General Principle can be realized in other embodiments without departing from the spirit or scope of the application.Therefore, the application It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one The widest scope of cause.

Claims (14)

1. a kind of unreasonable medical diagnosis on disease detection method applied to Clinical Decision Support Systems characterized by comprising
Obtain the medical record information and medical diagnosis on disease result of target object；
The object properties information and/or state of an illness information of the target object are obtained from the case information；
The determining and unmatched disease of object properties information from the medical diagnosis on disease result, and/or believe with the state of an illness Unmatched disease is ceased, as the unreasonable disease in the medical diagnosis on disease result.
2. the unreasonable medical diagnosis on disease detection method according to claim 1 applied to Clinical Decision Support Systems, special Sign is, further includes:
The unreasonable disease is deleted from the medical diagnosis on disease result；
Medical diagnosis on disease result after deleting the unreasonable disease medical diagnosis on disease result final as the target object.
3. the unreasonable medical diagnosis on disease detection method according to claim 1 applied to Clinical Decision Support Systems, special Sign is, determines and the unmatched disease of object properties information from the medical diagnosis on disease result, comprising:
Group's element is obtained from the object properties information of the target object based on multiple groups predetermined, wherein institute It states multiple groups and is based on object property definition；
Group belonging to the target object is determined from multiple groups predetermined based on group's element, as Target group；
The determining and unmatched disease of the target group from the medical diagnosis on disease result, as pair with the target object As the unmatched disease of attribute information.
4. the unreasonable medical diagnosis on disease detection method according to claim 3 applied to Clinical Decision Support Systems, special Sign is, described to determine and the unmatched disease of the target group from the medical diagnosis on disease result, comprising:
Subgraph relevant to the target group is determined from the full connexon set of graphs of group being obtained ahead of time, that determines is all Subgraph forms target subgraph set；Wherein, the full connexon set of graphs of the group is obtained based on the group relation figure constructed in advance, The group relation figure can characterize the relationship between multiple groups predetermined, the full connexon set of graphs of group In any subgraph in any node represent between a group, any two node have connection relationship；
Based on the target subgraph set, determined from the medical diagnosis on disease result and the unmatched disease of the target group Disease.
5. the unreasonable medical diagnosis on disease detection method according to claim 4 applied to Clinical Decision Support Systems, special Sign is, described to be based on the target subgraph set, is not determined from the medical diagnosis on disease result with the target group not The disease matched, comprising:
Obtain the corresponding probability of each subgraph in the target subgraph set and disease set, wherein any subgraph is corresponding general Rate passes through the determine the probability of each node in the subgraph, and the probability of any node is true by group relevant to node element It is fixed；
Based on the corresponding probability of subgraph each in the target subgraph set and disease set, target disease Making by Probability Sets is updated, Wherein, the target disease Making by Probability Sets includes the probability of each disease in target disease set, the target disease set packet Include disease involved in all case informations；
Based on updated target disease Making by Probability Sets, do not determined from the medical diagnosis on disease result with the target group not The disease matched.
6. the unreasonable medical diagnosis on disease detection method according to claim 5 applied to Clinical Decision Support Systems, special Sign is, described based on the corresponding probability of subgraph each in the target subgraph set and disease set, and it is general to update target disease Rate set, comprising:
The subgraph that acquisition one had not been obtained from the target subgraph set is as target subgraph；
Based on the corresponding disease set of the target subgraph, obtained from the group relation figure relevant to the target subgraph Subgraph, wherein include the corresponding disease of the target subgraph in the corresponding disease set of relevant to target subgraph subgraph Disease in set；
Based on the target subgraph and the corresponding probability of subgraph relevant to the target subgraph, the target disease is updated The probability of disease is specified in Making by Probability Sets, then executes the subgraph that obtains from the target subgraph set as target Figure, until the subgraph not had not been obtained in the target subgraph set, wherein the specified disease includes at least the target Disease in the corresponding disease set of subgraph.
7. the unreasonable medical diagnosis on disease detection method according to claim 5 applied to Clinical Decision Support Systems, special Sign is, described to be based on updated target disease Making by Probability Sets, determines from the medical diagnosis on disease result and the target The unmatched disease of group, comprising:
For any disease in the medical diagnosis on disease result:
The corresponding probability of the disease is obtained from the updated target disease Making by Probability Sets；
If the corresponding probability of the disease be less than predetermined probabilities threshold value, it is determined that the disease be and the unmatched disease of the target group Disease；
With obtain in the medical diagnosis on disease result with the unmatched disease of the target group.
8. being applied to the unreasonable medical diagnosis on disease of Clinical Decision Support Systems described according to claim 1~any one of 7 Detection method, which is characterized in that the determining and unmatched disease of state of an illness information from the medical diagnosis on disease result, comprising:
When the first object disease in the medical diagnosis on disease result comprising being related to position, the state of an illness letter based on the target object Breath determines the disease sites of the target object；
If the position that the disease sites of the target object and the first object disease are related to mismatches, it is determined that the disease The first object disease in diagnostic result is and the unmatched disease of state of an illness information.
9. the unreasonable medical diagnosis on disease detection method according to claim 8 applied to Clinical Decision Support Systems, special Sign is, determines whether the disease sites of the target object match with the position that the first object disease is related to, comprising:
Disease sites and the first object disease based on the portion structure tree and the target object pre-established are related to Position of the position in the portion structure tree, determine that the disease sites of the target object are related to the first object disease And position whether match, wherein the portion structure tree be position hyponymy tree.
10. being applied to the unreasonable medical diagnosis on disease of Clinical Decision Support Systems described according to claim 1~any one of 7 Detection method, which is characterized in that the determining and unmatched disease of state of an illness information from the medical diagnosis on disease result, comprising:
When in the medical diagnosis on disease result including the third target disease easily obscured with the second target disease, it is based on the target The state of an illness information of object determines the disease that the target object is suffered from；
If the disease that the target object is suffered from is second target disease, it is determined that described in the medical diagnosis on disease result Third target disease be and the unmatched disease of state of an illness information.
11. the unreasonable medical diagnosis on disease detection method according to claim 10 applied to Clinical Decision Support Systems, It is characterized in that, the state of an illness information based on the target object determines the disease that the target object is suffered from, comprising:
In the target disease disaggregated model that the state of an illness information input is pre-established, the disease that the target object is suffered from is obtained Disease；
Wherein, the target disease disaggregated model is that second target disease and the easy of third target disease composition are obscured Disease easily obscures classification of diseases model to corresponding, and the target disease disaggregated model uses disease for the second target disease It disease but is easily diagnosed as state of an illness sample and/or the disease of third target disease as the third target disease but is easily diagnosed as institute The state of an illness sample training for stating the second target disease obtains.
12. a kind of unreasonable medical diagnosis on disease detection device applied to Clinical Decision Support Systems characterized by comprising the One obtains module, the second acquisition module and unreasonable disease detection module；
Described first obtains module, for obtaining the medical record information and medical diagnosis on disease result of target object；
It is described second obtain module, for obtained from the case information target object object properties information and/or State of an illness information；
The unreasonable disease detection module, for from the medical diagnosis on disease result it is determining with the object properties information not The disease matched, and/or with the unmatched disease of state of an illness information, as the unreasonable disease in the medical diagnosis on disease result.
13. a kind of unreasonable medical diagnosis on disease detection device applied to Clinical Decision Support Systems characterized by comprising deposit Reservoir and processor；
The memory, for storing program；
The processor realizes determining as described in any one of claim 1~11 applied to clinic for executing described program Plan supports each step of the unreasonable medical diagnosis on disease detection method of system.
14. a kind of readable storage medium storing program for executing, is stored thereon with computer program, which is characterized in that the computer program is processed When device executes, the unreasonable disease applied to Clinical Decision Support Systems as described in any one of claim 1~11 is realized Each step of diagnosis detecting method.
CN201910328888.4A 2019-04-23 2019-04-23 Unreasonable disease diagnosis and detection method and device applied to clinical decision support system Active CN110033863B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910328888.4A CN110033863B (en) 2019-04-23 2019-04-23 Unreasonable disease diagnosis and detection method and device applied to clinical decision support system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910328888.4A CN110033863B (en) 2019-04-23 2019-04-23 Unreasonable disease diagnosis and detection method and device applied to clinical decision support system

Publications (2)

Publication Number Publication Date
CN110033863A true CN110033863A (en) 2019-07-19
CN110033863B CN110033863B (en) 2021-06-04

Family

ID=67239807

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910328888.4A Active CN110033863B (en) 2019-04-23 2019-04-23 Unreasonable disease diagnosis and detection method and device applied to clinical decision support system

Country Status (1)

Country Link
CN (1) CN110033863B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111081370A (en) * 2019-10-25 2020-04-28 中国科学院自动化研究所 User classification method and device
CN115148323A (en) * 2022-09-06 2022-10-04 北京鹰瞳科技发展股份有限公司 Apparatus, method and readable storage medium for disease prediction based on medical image
CN115147668A (en) * 2022-09-06 2022-10-04 北京鹰瞳科技发展股份有限公司 Training method of disease classification model, disease classification method and related products

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050108045A1 (en) * 2003-11-17 2005-05-19 Jiao Gong Method of network hospital of reclassification of diseases that have been made a diagnosis
CN104866727A (en) * 2015-06-02 2015-08-26 陈宽 Deep learning-based method for analyzing medical data and intelligent analyzer thereof
WO2015176027A1 (en) * 2014-05-15 2015-11-19 Globalvetlink, L.C. System and method for predicting effectiveness of animal treatments
CN105389465A (en) * 2015-10-29 2016-03-09 徐翼 Medical record data processing method, apparatus and system
CN107358014A (en) * 2016-11-02 2017-11-17 华南师范大学 The clinical pre-treating method and system of a kind of physiological data
CN107591201A (en) * 2017-09-13 2018-01-16 中国科学院重庆绿色智能技术研究院 A kind of dicision of diagnosis and treatment supports system
US20180101648A1 (en) * 2016-10-06 2018-04-12 Fujitsu Limited Computer apparatus and method to identify healthcare resources used by a patient of a medical institution
CN108182973A (en) * 2017-12-29 2018-06-19 湖南大学 A kind of Intelligent Diagnosis Technology of knowledge based collection of illustrative plates reasoning
CN108899070A (en) * 2018-05-31 2018-11-27 平安医疗科技有限公司 Prescription recommends generation method, device, computer equipment and storage medium
CN109102889A (en) * 2018-06-29 2018-12-28 平安科技（深圳）有限公司 Disease detection method, detection service device and computer readable storage medium

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050108045A1 (en) * 2003-11-17 2005-05-19 Jiao Gong Method of network hospital of reclassification of diseases that have been made a diagnosis
WO2015176027A1 (en) * 2014-05-15 2015-11-19 Globalvetlink, L.C. System and method for predicting effectiveness of animal treatments
CN104866727A (en) * 2015-06-02 2015-08-26 陈宽 Deep learning-based method for analyzing medical data and intelligent analyzer thereof
CN105389465A (en) * 2015-10-29 2016-03-09 徐翼 Medical record data processing method, apparatus and system
US20180101648A1 (en) * 2016-10-06 2018-04-12 Fujitsu Limited Computer apparatus and method to identify healthcare resources used by a patient of a medical institution
CN107358014A (en) * 2016-11-02 2017-11-17 华南师范大学 The clinical pre-treating method and system of a kind of physiological data
CN107591201A (en) * 2017-09-13 2018-01-16 中国科学院重庆绿色智能技术研究院 A kind of dicision of diagnosis and treatment supports system
CN108182973A (en) * 2017-12-29 2018-06-19 湖南大学 A kind of Intelligent Diagnosis Technology of knowledge based collection of illustrative plates reasoning
CN108899070A (en) * 2018-05-31 2018-11-27 平安医疗科技有限公司 Prescription recommends generation method, device, computer equipment and storage medium
CN109102889A (en) * 2018-06-29 2018-12-28 平安科技（深圳）有限公司 Disease detection method, detection service device and computer readable storage medium

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111081370A (en) * 2019-10-25 2020-04-28 中国科学院自动化研究所 User classification method and device
CN111081370B (en) * 2019-10-25 2023-11-03 中国科学院自动化研究所 User classification method and device
CN115148323A (en) * 2022-09-06 2022-10-04 北京鹰瞳科技发展股份有限公司 Apparatus, method and readable storage medium for disease prediction based on medical image
CN115147668A (en) * 2022-09-06 2022-10-04 北京鹰瞳科技发展股份有限公司 Training method of disease classification model, disease classification method and related products

Also Published As

Publication number Publication date
CN110033863B (en) 2021-06-04

Similar Documents

Publication Publication Date Title
US9668702B2 (en) Systems, methods, and computer readable media for using descriptors to identify when a subject is likely to have a dysmorphic feature
CN110033863A (en) Applied to the unreasonable medical diagnosis on disease detection method of Clinical Decision Support Systems, device
CN108766561B (en) Method, apparatus, computer device and storage medium for processing disease information
Zhao et al. Retinal artery and vein classification via dominant sets clustering-based vascular topology estimation
CN110675944A (en) Triage method and device, computer equipment and medium
US20160042134A1 (en) Method of calculating a score of a medical suggestion as a support in medical decision making
JP7082993B2 (en) Medical image processing equipment, methods and programs, diagnostic support equipment, methods and programs, and medical support systems and methods
KR101850772B1 (en) Method and apparatus for managing clinical meta database
US20130231953A1 (en) Method, system and computer program product for aggregating population data
JP2023513060A (en) instrument tracking machine
CN106599553A (en) Disease early-warning method and device
JP7033202B2 (en) Medical image processing equipment and methods, machine learning systems, programs and storage media
CN109191451A (en) Method for detecting abnormality, device, equipment and medium
JP2022036125A (en) Contextual filtering of examination values
CN107145715A (en) A kind of clinical medical intelligent discriminating gear based on election algorithm
CN107978343A (en) The evaluation system of electronic health record
CN115438040A (en) Pathological archive information management method and system
CN109997201A (en) For the accurate clinical decision support using data-driven method of plurality of medical knowledge module
Jiang et al. Improving the generalizability of infantile cataracts detection via deep learning-based lens partition strategy and multicenter datasets
WO2022205779A1 (en) Processing method and apparatus based on multi-modal eye detection data, and terminal device
CN110490138A (en) A kind of data processing method and device, storage medium, electronic equipment
CN113707304A (en) Triage data processing method, device, equipment and storage medium
CN108986918A (en) Aided diagnosis method, device, terminal device and computer storage medium
Hong et al. A novel hierarchical deep learning framework for diagnosing multiple visual impairment diseases in the clinical environment
WO2022141929A1 (en) Health self-test system, server, and health testing system

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