CN112151182A - Intelligent medical advice generation method and system based on greedy search - Google Patents
Intelligent medical advice generation method and system based on greedy search Download PDFInfo
- Publication number
- CN112151182A CN112151182A CN202010817792.7A CN202010817792A CN112151182A CN 112151182 A CN112151182 A CN 112151182A CN 202010817792 A CN202010817792 A CN 202010817792A CN 112151182 A CN112151182 A CN 112151182A
- Authority
- CN
- China
- Prior art keywords
- index
- patient
- physiological
- adjustment
- physiological index
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013136 deep learning model Methods 0.000 claims abstract description 13
- 230000000007 visual effect Effects 0.000 claims abstract description 7
- 230000007246 mechanism Effects 0.000 claims description 36
- 230000009191 jumping Effects 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 230000001174 ascending effect Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 description 4
- 238000012800 visualization Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 208000017667 Chronic Disease Diseases 0.000 description 1
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 229910002092 carbon dioxide Inorganic materials 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003828 downregulation Effects 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 229910052700 potassium Inorganic materials 0.000 description 1
- 239000011591 potassium Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012502 risk assessment Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000003827 upregulation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- 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
-
- 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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses an intelligent medical advice generation method and system based on greedy search, wherein the method comprises the following steps: s100, inputting the electronic medical record of the patient into a trained deep learning model, and acquiring an important index and a death risk index of the physiological index of the patient in each visit; s200, carrying out greedy search on the important index and the death risk index of the physiological index of each visit of the patient to obtain an adjustment scheme of each physiological index of the patient, and generating an intelligent medical suggestion for the patient; s300, displaying the intelligent medical advice based on the visual interface. The invention can excavate more valuable information with reference meaning from the patient electronic medical record data of the hospital to help the doctor to diagnose.
Description
Technical Field
The invention relates to the technical field of intelligent medical treatment, in particular to an intelligent medical treatment suggestion generation method and system based on greedy search.
Background
For the terminal chronic disease patients and critical emergency patients, the death risk of the patients is evaluated, which is helpful for helping medical personnel to take targeted treatment measures as soon as possible and improve the survival probability of the patients. However, the death risk assessment is very complex, and needs to comprehensively analyze multiple physiological indexes of the patient, consider static information of the age, sex, past medical history and the like of the patient, and make a final judgment in combination with a large amount of medical knowledge, and only doctors with abundant medical knowledge and clinical experience can obtain a more accurate assessment result on the premise of investing a large amount of time and energy. In clinical practice, the medical institution often has poor evaluation effect due to lack of sufficient expert resources. On the other hand, in recent years, various medical institutions have accumulated a large number of items including: electronic medical record data including basic information of patients, examination and examination reports, medication and treatment records, and the like.
Under the background, the invention designs an intelligent medical advice generation technical scheme based on electronic medical record data. The death risk index of the patient and the importance degree of each physiological index in each visit are obtained by using the deep learning model proposed in the reference, and on the basis, an intelligent medical advice is generated for the specific patient to assist a doctor in diagnosis.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an intelligent medical advice generation method and system based on greedy search, which can be used for mining more valuable and referential information from electronic medical record data of patients in hospitals to help doctors to diagnose.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
an intelligent medical advice generation method based on greedy search, the method comprising:
(1) inputting the electronic medical record of the patient into a trained deep learning model, and acquiring an important index and a death risk index of the physiological index of each visit of the patient;
(2) greedy search is carried out on the important index and the death risk index of the physiological index of each visit of the patient, the adjustment scheme of each physiological index of the patient is obtained, and an intelligent medical suggestion for the patient is generated;
(3) and displaying the intelligent medical advice based on a visual interface.
Further, the method for generating an intelligent medical advice based on greedy search as described above, step (2) includes:
(2.1) selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
(2.2) obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
(2.3) adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
(2.4) recalculating the significance index and the death risk index of the physiological index of each visit of the patient using the adjusted data;
(2.5) judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is less than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
(2.6) if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and skipping to the step (2.1);
(2.7) if not, judging whether the current physiological index meets the condition of a rollback mechanism, if so, performing rollback operation, and jumping to the step (2.1);
(2.8) if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, skipping to the step (2.2), and if not, skipping to the step (2.1).
Further, the method for generating intelligent medical advice based on greedy search as described above, wherein the step (2.1) includes:
and eliminating the physiological indexes which are just rolled back and the physiological indexes in the shielding set from all the physiological indexes of the patient, and selecting the physiological index with the largest importance index from the rest physiological indexes of the patient as the current physiological index to be adjusted.
Further, the method for generating intelligent medical advice based on greedy search as described above, wherein the step (2.2) includes:
if the current physiological index is adjusted for the first time, setting the adjustment proportion set by the current physiological index as a preset adjustment proportion, and setting the adjustment direction as ascending, namely increasing the numerical value of the current physiological index by a preset percentage;
if the current physiological index is adjusted for the second time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and keeping the adjustment proportion unchanged;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and setting the adjustment proportion to be half of the previous adjustment proportion;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is reduced, the adjustment direction and the adjustment proportion of the current physiological index are kept unchanged.
Further, the method for generating intelligent medical advice based on greedy search as described above, wherein the step (2.6) includes:
if the current physiological index meets the following two conditions, the current physiological index is added into the shielding set and is not adjusted any more:
a. the value of the current physiological index exceeds the normal range of the value of the physiological index and tries to continuously adjust towards the direction deviating from the normal range;
b. the current physiological index has been adjusted a preset number of times without the patient's mortality risk value falling too much.
Further, the method for generating intelligent medical advice based on greedy search as described above, wherein the step (2.7) includes:
if the current physiological index meets the following conditions, setting the value of the physiological index, the recorded current death risk value, the current important index of the physiological index, the recorded death risk value before adjustment and the important index before adjustment of the physiological index as the values before the previous adjustment:
the current physiological indicators exhibit an abnormal effect on the mortality risk index of the patient.
Further, the method for generating an intelligent medical advice based on greedy search as described above, step (2) includes:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
An intelligent medical advice generation system based on greedy search, the system comprising:
the acquisition module is used for inputting the electronic medical record of the patient into the trained deep learning model and acquiring the important index and death risk index of the physiological index of the patient for each visit;
the generation module is used for carrying out greedy search on the important index and the death risk index of the physiological index of each visit of the patient to obtain an adjustment scheme of each physiological index of the patient and generate an intelligent medical suggestion for the patient;
and the display module is used for displaying the intelligent medical advice based on a visual interface.
Further, an intelligent medical advice generation system based on greedy search as described above, the generation module is configured to:
(2.1) selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
(2.2) obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
(2.3) adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
(2.4) recalculating the significance index and the death risk index of the physiological index of each visit of the patient using the adjusted data;
(2.5) judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is less than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
(2.6) if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and skipping to the step (2.1);
(2.7) if not, judging whether the current physiological index meets the condition of a rollback mechanism, if so, performing rollback operation, and jumping to the step (2.1);
(2.8) if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, skipping to the step (2.2), and if not, skipping to the step (2.1).
Further, an intelligent medical advice generation system based on greedy search as described above, the generation module is further configured to:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
The invention has the beneficial effects that: the method and the system provided by the invention can be used for early warning a high-risk patient, prompting physiological indexes needing important attention, giving corresponding medical suggestions, and guiding a doctor to perform simulation intervention on the physiological indexes and observe possible future illness states of the patient, so that the body condition trend of the patient is controlled to be developed as much as possible.
Drawings
Fig. 1 is a schematic flow chart of an intelligent medical advice generation method based on greedy search according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of an intelligent medical advice generation method based on greedy search according to an embodiment of the present invention;
FIG. 3 is a flowchart for obtaining a patient mortality risk index and a physiological index significance index provided in an embodiment of the present invention;
fig. 4 is a flowchart of an intelligent medical advice generation method based on greedy search according to an embodiment of the present invention;
fig. 5 is a schematic diagram of an intelligent medical advice visualization interface provided in an embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and the detailed description.
Reference to the literature
Ma L,Zhang C,Wang Y,et al.Concare:Personalized clinical feature embedding via capturing the healthcare context[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020,34(01):833-840.
Interpretation of terms
Electronic medical record
Electronic Medical Records (EMR) are also known as computerized Medical Record systems or computer-based patient records. It is a digitalized medical record stored, managed, transmitted and reproduced by electronic equipment (computer, health card, etc.) to replace the hand-written paper case history. Its contents include all the information of the paper case history.
Vital index of physiological index
The electronic medical record data of the previous n times of patient visits are input into the deep learning model provided by reference, and the weighting parameters (the sum is 1) of the death prediction result of each physiological index of each patient visit in the last year can be obtained. The weighting parameter of each physiological index multiplied by 100 is defined as the importance index of the physiological index, the range is 0-100, and the larger the value is, the more important the physiological index is.
Mortality risk index
A mortality risk index, output from the deep learning model proposed in the literature references, is an indicator (ranging from 0 to 100) used to assess the risk of mortality in a patient in the next year, with greater than 50 indicating a more severe condition in the patient.
Patient health status presentation learning
Predicting a patient's clinical outcome from an electronic medical record is a fundamental research topic of medical informatics. The reference proposes a deep learning model for predicting the death probability of a patient within one year by combining basic information of the patient and a test examination report, and the prediction process is a learning process expressed by the health state of the patient.
As shown in fig. 1, an intelligent medical advice generation method based on greedy search includes:
s100, inputting the electronic medical record of the patient into a trained deep learning model, and acquiring an important index and a death risk index of the physiological index of the patient in each visit;
s200, carrying out greedy search on the important index and the death risk index of the physiological index of each visit of the patient to obtain an adjustment scheme of each physiological index of the patient, and generating an intelligent medical suggestion for the patient;
s300, displaying the intelligent medical advice based on the visual interface.
The step S200 includes:
s210, selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
s220, obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
s230, adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
s240, recalculating the important index and death risk index of the physiological index of each visit of the patient by using the adjusted data;
s250, judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is smaller than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
s260, if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and jumping to S210;
s270, if not, judging whether the current physiological indexes meet the condition of a rollback mechanism, if so, performing rollback operation, and jumping to S210;
and S280, if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, jumping to S220, and if not, jumping to S210.
Step S210 includes:
and eliminating the physiological indexes which are just rolled back and the physiological indexes in the shielding set from all the physiological indexes of the patient, and selecting the physiological index with the largest importance index from the rest physiological indexes of the patient as the current physiological index to be adjusted.
Step S220 includes:
if the current physiological index is adjusted for the first time, setting the adjustment proportion set by the current physiological index as a preset adjustment proportion, and setting the adjustment direction as ascending, namely increasing the numerical value of the current physiological index by a preset percentage;
if the current physiological index is adjusted for the second time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and keeping the adjustment proportion unchanged;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and setting the adjustment proportion to be half of the previous adjustment proportion;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is reduced, the adjustment direction and the adjustment proportion of the current physiological index are kept unchanged.
The step S260 includes:
if the current physiological index meets the following two conditions, the current physiological index is added into the shielding set and is not adjusted any more:
a. the value of the current physiological index exceeds the normal range of the value of the physiological index and tries to continuously adjust towards the direction deviating from the normal range;
b. the current physiological index has been adjusted a preset number of times without the patient's mortality risk value falling too much.
Step S270 includes:
if the current physiological index meets the following conditions, setting the value of the physiological index, the recorded current death risk value, the current important index of the physiological index, the recorded death risk value before adjustment and the important index before adjustment of the physiological index as the values before the previous adjustment:
the current physiological indicators exhibit an abnormal effect on the mortality risk index of the patient.
The step S200 includes:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
As shown in fig. 2, the method of the present invention comprises three main steps: (1) acquiring important indexes of physiological indexes and death risk indexes of patients in each visit; (2) intelligent medical advice generation; (3) visualization and interface element design.
(1) Obtaining important indexes of physiological indexes and death risk indexes of patients at each visit
As shown in fig. 3, the importance index and death risk index of the physiological index of each visit of the patient can be obtained by inputting the electronic medical record of the patient (including the basic information of the patient and the physiological index data of each visit) into the deep learning model proposed in the reference.
(2) Intelligent medical advice generation
Based on the deep learning model provided in the reference literature, the important index of the physiological index and the death risk index of each visit of the patient can be obtained, the death risk index of the patient is subjected to deep search, and a preset death risk index target can be reached, so that a plurality of layers of medical advice are generated.
As shown in fig. 4, an intelligent medical advice generation flow. Inputting: maximum iteration times, target death risk index, electronic medical record data of each visit of the patient; and (3) outputting: adjusting the physiological index of the patient.
The termination condition of the process is that the adjustment times of the physiological indexes reach the maximum iteration times or the death risk value after the adjustment of the patient is less than or equal to the target death risk value. In order to generate an intelligent medical suggestion, a loop iteration mode is adopted to adjust the numerical value of each physiological index of the last visit of a patient according to a certain proportion and direction, and a rollback mechanism, a shielding mechanism, a direction and scale adjustment mechanism and a mechanism for selecting the physiological index to be adjusted are added.
1) Main body circulation flow path
S1, selecting indexes according to a to-be-adjusted physiological index selection mechanism;
s2, obtaining the adjusted direction and proportion according to the adjusted direction and scale conversion mechanism;
s3, adjusting the value of the physiological index of the patient in the last visit according to the obtained direction and proportion;
s4, recalculating the importance index and death risk index of the physiological indexes of each visit of the patient by using the adjusted data;
s5, if the adjustment times reach the maximum iteration times or the death risk value after the adjustment of the patient is less than or equal to the target death risk value, terminating the function and outputting an adjustment scheme;
s6, judging a shielding mechanism and a rollback mechanism, and jumping to S1 if the conditions are met;
and S7, if the current index is adjusted and is the index with the highest importance index, jumping to S2, and otherwise, jumping to S1.
2) Shielding mechanism
The physiological indexes meeting the following two conditions are added into a shielding set, and are not adjusted in the whole intelligent suggestion generation process.
1. The value of the current physiological indicator has exceeded the normal range of values for the physiological indicator and attempts to continue to adjust in a direction away from the normal range.
2. There have been 5 attempts to adjust, but none of the patients' mortality risk values have declined.
3) Rollback mechanism
The rollback operation specifically refers to setting all the numerical value of a certain physiological index, the recorded current death risk value, the current importance index of the index, the recorded death risk value before adjustment and the importance index before index adjustment as the numerical value before previous adjustment. The rollback conditions that need to be met are as follows:
the currently adjusted physiological index exhibits an abnormal effect on the mortality risk index of the patient.
4) Adjusting direction and scale transformation mechanism
For a certain physiological index of the patient:
a. if the physiological index is adjusted for the first time, the adjustment proportion is set to be 3%, the adjustment direction is increased, namely, the value of the physiological index is increased by a certain proportion.
b. If the importance index of the physiological index is increased after the second adjustment and the last adjustment, the adjustment direction is set to be opposite to the last adjustment, and the adjustment proportion is unchanged.
c. If the physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is increased, the adjusting direction is set to be opposite to the adjusting direction of the last time, and the adjusting proportion is set to be half of the adjusting proportion of the last time.
d. If the importance index of the physiological index is not adjusted for the first time and is adjusted for the last time and is reduced, the adjusting direction and the adjusting proportion are kept unchanged.
5) Mechanism for selecting physiological index to be adjusted
Firstly, the physiological indexes which are just rolled back and the physiological indexes in the shielding set are removed from all the physiological indexes of the patient, and then the physiological index with the largest importance index is selected from the rest physiological indexes to be used as the current physiological index to be adjusted.
(3) Visualization and interface element design
The visualization interface shown in fig. 5 first gives the time of the last visit of the patient, indicating that a recommendation is made to the physician at that time. The two most significant indicators for the patient at the current time are then presented to the physician (carbon dioxide binding and blood potassium in figure 5). And then giving two grades of intelligent medical advice of doctors, wherein the first grade is an index for advising adjustment if the death risk index is reduced to be below 40, and the second grade is an index for advising adjustment for reducing the death risk index to the maximum extent. If the patient's own mortality risk index is below 40, no first-gear advice is displayed.
In the specific suggestion part, the physiological index to be adjusted, the original value of the physiological index, the adjusted value, the adjusted proportion and the adjusted value are presented. In order to make the user more clearly see, each physiological index noun may be assigned a specific color, and the up-regulation proportion is emphasized by red and the down-regulation proportion by green.
The invention excavates more valuable and referential information from the electronic medical record data of the patients in the hospital to help doctors to diagnose. The early warning can be carried out on a high-risk patient, the physiological indexes needing important attention are prompted, corresponding medical suggestions are given, and a doctor is guided to carry out simulation intervention on the physiological indexes and observe possible future illness states of the patient, so that the body condition trend of the patient is controlled to be developed as well as possible.
The invention also provides an intelligent medical advice generating system based on greedy search, which comprises:
the acquisition module is used for inputting the electronic medical record of the patient into the trained deep learning model and acquiring the important index and death risk index of the physiological index of the patient for each visit;
the generation module is used for carrying out greedy search on the important index and the death risk index of the physiological index of each visit of the patient to obtain an adjustment scheme of each physiological index of the patient and generate an intelligent medical suggestion for the patient;
and the display module is used for displaying the intelligent medical advice based on the visual interface.
The generation module is to:
(2.1) selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
(2.2) obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
(2.3) adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
(2.4) recalculating the importance index and death risk index of the physiological index of each visit of the patient by using the adjusted data;
(2.5) judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is less than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
(2.6) if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and skipping to the step (2.1);
(2.7) if not, judging whether the current physiological index meets the condition of a rollback mechanism, if so, performing rollback operation, and jumping to the step (2.1);
(2.8) if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, skipping to the step (2.2), and if not, skipping to the step (2.1).
The step (2.1) comprises the following steps:
and eliminating the physiological indexes which are just rolled back and the physiological indexes in the shielding set from all the physiological indexes of the patient, and selecting the physiological index with the largest importance index from the rest physiological indexes of the patient as the current physiological index to be adjusted.
The step (2.2) comprises the following steps:
if the current physiological index is adjusted for the first time, setting the adjustment proportion set by the current physiological index as a preset adjustment proportion, and setting the adjustment direction as ascending, namely increasing the numerical value of the current physiological index by a preset percentage;
if the current physiological index is adjusted for the second time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and keeping the adjustment proportion unchanged;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and setting the adjustment proportion to be half of the previous adjustment proportion;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is reduced, the adjustment direction and the adjustment proportion of the current physiological index are kept unchanged.
The step (2.6) comprises the following steps:
if the current physiological index meets the following two conditions, the current physiological index is added into the shielding set and is not adjusted any more:
a. the value of the current physiological index exceeds the normal range of the value of the physiological index and tries to continuously adjust towards the direction deviating from the normal range;
b. the current physiological index has been adjusted a preset number of times without the patient's mortality risk value falling too much.
The step (2.7) comprises the following steps:
if the current physiological index meets the following conditions, setting the value of the physiological index, the recorded current death risk value, the current important index of the physiological index, the recorded death risk value before adjustment and the important index before adjustment of the physiological index as the values before the previous adjustment:
the current physiological indicators exhibit an abnormal effect on the mortality risk index of the patient.
The generation module is further to:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is intended to include such modifications and variations.
Claims (10)
1. An intelligent medical advice generation method based on greedy search, the method comprising:
(1) inputting the electronic medical record of the patient into a trained deep learning model, and acquiring an important index and a death risk index of the physiological index of each visit of the patient;
(2) greedy search is carried out on the important index and the death risk index of the physiological index of each visit of the patient, the adjustment scheme of each physiological index of the patient is obtained, and an intelligent medical suggestion for the patient is generated;
(3) and displaying the intelligent medical advice based on a visual interface.
2. The method for generating intelligent medical advice based on greedy search according to claim 1, wherein the step (2) comprises:
(2.1) selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
(2.2) obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
(2.3) adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
(2.4) recalculating the significance index and the death risk index of the physiological index of each visit of the patient using the adjusted data;
(2.5) judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is less than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
(2.6) if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and skipping to the step (2.1);
(2.7) if not, judging whether the current physiological index meets the condition of a rollback mechanism, if so, performing rollback operation, and jumping to the step (2.1);
(2.8) if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, skipping to the step (2.2), and if not, skipping to the step (2.1).
3. The greedy search based intelligent medical advice generation method according to claim 2, wherein the step (2.1) comprises:
and eliminating the physiological indexes which are just rolled back and the physiological indexes in the shielding set from all the physiological indexes of the patient, and selecting the physiological index with the largest importance index from the rest physiological indexes of the patient as the current physiological index to be adjusted.
4. The method of claim 2, wherein the step (2.2) comprises:
if the current physiological index is adjusted for the first time, setting the adjustment proportion set by the current physiological index as a preset adjustment proportion, and setting the adjustment direction as ascending, namely increasing the numerical value of the current physiological index by a preset percentage;
if the current physiological index is adjusted for the second time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and keeping the adjustment proportion unchanged;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is increased, setting the adjustment direction of the current physiological index to be the opposite direction of the previous adjustment, and setting the adjustment proportion to be half of the previous adjustment proportion;
if the current physiological index is not adjusted for the first time and the importance index of the physiological index adjusted for the last time is reduced, the adjustment direction and the adjustment proportion of the current physiological index are kept unchanged.
5. The method of claim 2, wherein the step (2.6) comprises:
if the current physiological index meets the following two conditions, the current physiological index is added into the shielding set and is not adjusted any more:
a. the value of the current physiological index exceeds the normal range of the value of the physiological index and tries to continuously adjust towards the direction deviating from the normal range;
b. the current physiological index has been adjusted a preset number of times without the patient's mortality risk value falling too much.
6. The method of claim 2, wherein the step (2.7) comprises:
if the current physiological index meets the following conditions, setting the value of the physiological index, the recorded current death risk value, the current important index of the physiological index, the recorded death risk value before adjustment and the important index before adjustment of the physiological index as the values before the previous adjustment:
the current physiological indicators exhibit an abnormal effect on the mortality risk index of the patient.
7. The method for generating intelligent medical advice based on greedy search according to claim 1, wherein the step (2) comprises:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
8. An intelligent medical advice generation system based on greedy search, the system comprising:
the acquisition module is used for inputting the electronic medical record of the patient into the trained deep learning model and acquiring the important index and death risk index of the physiological index of the patient for each visit;
the generation module is used for carrying out greedy search on the important index and the death risk index of the physiological index of each visit of the patient to obtain an adjustment scheme of each physiological index of the patient and generate an intelligent medical suggestion for the patient;
and the display module is used for displaying the intelligent medical advice based on a visual interface.
9. The system of claim 8, wherein the generation module is configured to:
(2.1) selecting the physiological index with the largest importance index from the currently selectable physiological indexes of the patient as the current physiological index according to a mechanism for selecting the physiological index to be adjusted;
(2.2) obtaining the adjustment direction and the adjustment proportion of the current physiological index according to the adjustment direction and the scale conversion mechanism;
(2.3) adjusting the value of the current physiological index of the patient in the last visit according to the adjustment direction and the adjustment proportion;
(2.4) recalculating the significance index and the death risk index of the physiological index of each visit of the patient using the adjusted data;
(2.5) judging whether the adjustment times of the current physiological indexes reach the preset maximum iteration times or whether the death risk value of the patient after adjustment is less than or equal to the preset target death risk value or not, and if so, outputting the adjustment scheme of the current physiological indexes;
(2.6) if not, judging whether the current physiological index meets the condition of a shielding mechanism, if so, carrying out shielding operation, and skipping to the step (2.1);
(2.7) if not, judging whether the current physiological index meets the condition of a rollback mechanism, if so, performing rollback operation, and jumping to the step (2.1);
(2.8) if not, judging whether the adjusted current physiological index is the physiological index with the maximum importance index, if so, skipping to the step (2.2), and if not, skipping to the step (2.1).
10. The system of claim 8, wherein the generation module is further configured to:
generating an intelligent medical advice according to the adjustment scheme of each physiological index of the patient, wherein the intelligent medical advice comprises: reducing the patient's mortality risk index below a predetermined value suggests an adjusted physiological indicator, and minimizing the patient's mortality risk index suggests an adjusted physiological indicator.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010817792.7A CN112151182A (en) | 2020-08-14 | 2020-08-14 | Intelligent medical advice generation method and system based on greedy search |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010817792.7A CN112151182A (en) | 2020-08-14 | 2020-08-14 | Intelligent medical advice generation method and system based on greedy search |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112151182A true CN112151182A (en) | 2020-12-29 |
Family
ID=73888027
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010817792.7A Pending CN112151182A (en) | 2020-08-14 | 2020-08-14 | Intelligent medical advice generation method and system based on greedy search |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112151182A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1640888A2 (en) * | 2004-09-08 | 2006-03-29 | IpoCare GmbH & Co. KG | Method for estimating and supervising the medical risks of health problems for a patient |
US20100057490A1 (en) * | 2006-05-30 | 2010-03-04 | The University Of North Carolina At Chapel Hill | Methods, systems, and computer program products for evaluating a patient in a pediatric intensive care unit |
WO2012006174A2 (en) * | 2010-06-29 | 2012-01-12 | The University Of North Carolina At Chapel Hill | Methods, systems, and computer readable media for evaluating a hospital patient's risk of mortality |
CN102567624A (en) * | 2011-12-07 | 2012-07-11 | 南京毗邻医疗科技有限公司 | Individual intelligent medical service system on basis of diagnosis and treatment templates and illness state templates |
CN108172290A (en) * | 2017-12-12 | 2018-06-15 | 昆明亿尚科技有限公司 | A kind of hepatopath's nutrition in postoperative suggesting method based on artificial intelligence technology |
CN109310318A (en) * | 2015-12-16 | 2019-02-05 | 威斯康星州医药大学股份有限公司 | System and method for quantitatively being characterized to Alzheimer's disease risk case based on multi-modal biomarkcr data |
CN110111896A (en) * | 2019-05-20 | 2019-08-09 | 卫宁健康科技集团股份有限公司 | The recognition methods and system of medical-risk |
-
2020
- 2020-08-14 CN CN202010817792.7A patent/CN112151182A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1640888A2 (en) * | 2004-09-08 | 2006-03-29 | IpoCare GmbH & Co. KG | Method for estimating and supervising the medical risks of health problems for a patient |
US20100057490A1 (en) * | 2006-05-30 | 2010-03-04 | The University Of North Carolina At Chapel Hill | Methods, systems, and computer program products for evaluating a patient in a pediatric intensive care unit |
WO2012006174A2 (en) * | 2010-06-29 | 2012-01-12 | The University Of North Carolina At Chapel Hill | Methods, systems, and computer readable media for evaluating a hospital patient's risk of mortality |
CN102567624A (en) * | 2011-12-07 | 2012-07-11 | 南京毗邻医疗科技有限公司 | Individual intelligent medical service system on basis of diagnosis and treatment templates and illness state templates |
CN109310318A (en) * | 2015-12-16 | 2019-02-05 | 威斯康星州医药大学股份有限公司 | System and method for quantitatively being characterized to Alzheimer's disease risk case based on multi-modal biomarkcr data |
CN108172290A (en) * | 2017-12-12 | 2018-06-15 | 昆明亿尚科技有限公司 | A kind of hepatopath's nutrition in postoperative suggesting method based on artificial intelligence technology |
CN110111896A (en) * | 2019-05-20 | 2019-08-09 | 卫宁健康科技集团股份有限公司 | The recognition methods and system of medical-risk |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220330823A1 (en) | Systems and devices for telemetry monitoring management | |
US8670997B2 (en) | Quality metric extraction and editing for medical data | |
US20190311809A1 (en) | Patient status monitor and method of monitoring patient status | |
US20090094053A1 (en) | Diagnosis through graphical representation of patient characteristics | |
WO2022099668A1 (en) | Method and system for precise health management and risk early warning based on association between familial genetic disease and sign data | |
JP2014520335A (en) | Evaluation system and evaluation method | |
WO2009083886A1 (en) | Presenting patient relevant studies for clinical decision making | |
Kang et al. | Identifying nurses’ concern concepts about patient deterioration using a standard nursing terminology | |
US20210043326A1 (en) | Methods and systems for telemetry monitoring protocol management | |
KR20150007468A (en) | Clinical Decision Support System and Device supporting the same | |
CN116936134B (en) | Complications monitoring method and system based on nursing morning shift data | |
CN114613487A (en) | Pregnancy decision support system and method | |
CN112205965B (en) | Health risk key event detection method and system based on time window cutting | |
Fortier et al. | Development of a mobile cardiac wellness application and integrated wearable sensor suite | |
JP2021039683A (en) | Medical care assistance method, medical care assisting system, learning model generation method, and medical care assisting program | |
CN112151182A (en) | Intelligent medical advice generation method and system based on greedy search | |
Lin et al. | Development of a telehealthcare decision support system for patients discharged from the hospital | |
CN111317440A (en) | Early warning method for patient, monitoring device using the method and readable storage medium | |
Shankaracharya et al. | Java-based diabetes type 2 prediction tool for better diagnosis | |
JPH11232359A (en) | Method and system for processing at-home blood test result | |
Zavyalova et al. | Designing a mobile recommender system for treatment adherence improvement among hypertensives | |
JP2023020667A (en) | Medical information processing method, medical information processing device, and program | |
Gupta et al. | An overview of clinical decision support system (CDSS) as a computational tool and its applications in public health | |
Pustozerov et al. | Health monitoring system for patients with gestational diabetes mellitus based on nutrition diaries and fitness bracelets | |
CN113140323A (en) | Health portrait generation method, system, medium and server |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201229 |
|
RJ01 | Rejection of invention patent application after publication |