CN117995347A - Medical record content quality control method and device, electronic equipment and storage medium - Google Patents
Medical record content quality control method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN117995347A CN117995347A CN202410406684.9A CN202410406684A CN117995347A CN 117995347 A CN117995347 A CN 117995347A CN 202410406684 A CN202410406684 A CN 202410406684A CN 117995347 A CN117995347 A CN 117995347A
- Authority
- CN
- China
- Prior art keywords
- text
- question
- language model
- medical
- answer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003908 quality control method Methods 0.000 title claims abstract description 103
- 238000012549 training Methods 0.000 claims abstract description 107
- 238000003745 diagnosis Methods 0.000 claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 43
- 238000004590 computer program Methods 0.000 claims description 20
- 238000007781 pre-processing Methods 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 11
- 238000004140 cleaning Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 238000007405 data analysis Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 6
- 208000014674 injury Diseases 0.000 description 6
- 206010011224 Cough Diseases 0.000 description 4
- 201000004331 Henoch-Schoenlein purpura Diseases 0.000 description 4
- 206010019617 Henoch-Schonlein purpura Diseases 0.000 description 4
- 208000031814 IgA Vasculitis Diseases 0.000 description 4
- 208000036626 Mental retardation Diseases 0.000 description 4
- 206010037660 Pyrexia Diseases 0.000 description 4
- 206010039580 Scar Diseases 0.000 description 4
- 206010047700 Vomiting Diseases 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 4
- 238000002512 chemotherapy Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 208000015446 immunoglobulin a vasculitis Diseases 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000001356 surgical procedure Methods 0.000 description 4
- 230000008733 trauma Effects 0.000 description 4
- 206010053942 Cerebral haematoma Diseases 0.000 description 3
- 241000202934 Mycoplasma pneumoniae Species 0.000 description 3
- 208000002193 Pain Diseases 0.000 description 3
- 208000002667 Subdural Hematoma Diseases 0.000 description 3
- 208000002847 Surgical Wound Diseases 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000007428 craniotomy Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 238000002695 general anesthesia Methods 0.000 description 3
- 210000003141 lower extremity Anatomy 0.000 description 3
- 230000036407 pain Effects 0.000 description 3
- 230000008673 vomiting Effects 0.000 description 3
- 206010018985 Haemorrhage intracranial Diseases 0.000 description 2
- 208000008574 Intracranial Hemorrhages Diseases 0.000 description 2
- 206010028980 Neoplasm Diseases 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 2
- 201000011510 cancer Diseases 0.000 description 2
- 238000009535 clinical urine test Methods 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 230000008961 swelling Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 208000004998 Abdominal Pain Diseases 0.000 description 1
- 208000032544 Cicatrix Diseases 0.000 description 1
- 206010023232 Joint swelling Diseases 0.000 description 1
- 208000000913 Kidney Calculi Diseases 0.000 description 1
- 201000008235 Mycoplasma pneumoniae pneumonia Diseases 0.000 description 1
- 206010029148 Nephrolithiasis Diseases 0.000 description 1
- 206010035724 Pneumonia mycoplasmal Diseases 0.000 description 1
- 206010037549 Purpura Diseases 0.000 description 1
- 241001672981 Purpura Species 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 210000003414 extremity Anatomy 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000002483 medication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000474 nursing effect Effects 0.000 description 1
- 231100000241 scar Toxicity 0.000 description 1
- 230000037387 scars Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 230000002485 urinary effect Effects 0.000 description 1
- 210000002700 urine Anatomy 0.000 description 1
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
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- 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)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Pathology (AREA)
- Computing Systems (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention relates to a medical record content quality control method, a device, electronic equipment and a storage medium, wherein the method comprises the following steps: medical text data is acquired and analyzed to convert the medical text data into a first question text and a first answer text corresponding to the first question text. And taking a plurality of different first question texts as input of the large language model, and taking the first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training. And receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text. The method can complete the medical record content quality control rule which needs to be exhausted by using the context learning capability of the large language model, and only needs to give training samples to conduct context learning in the reasoning stage, so that the quality control effect can be ensured when the diagnosis quantity is large.
Description
Technical Field
The invention relates to the technical field of medical assistance, in particular to a medical record content quality control method, a medical record content quality control device, electronic equipment and a storage medium.
Background
The medical record quality control has the significance of ensuring standard, accurate and complete medical record writing, is used for improving medical quality, ensuring patient safety and being beneficial to clinical scientific research and medical education. Medical record quality control is an important link in daily management work of medical institutions at present, and is also one of important marks for evaluating medical quality.
Firstly, medical record quality control can ensure medical record writing standard, because medical record is a detailed record of patient's condition and is also the basis for doctor to diagnose, treat and manage diseases. The nonstandard part in the medical record writing, such as nonstandard writing format, inaccurate diagnosis, unreasonable treatment scheme and the like, can be timely found and corrected through medical record quality control, so that the standardization of the medical record writing needs to be ensured. Secondly, medical quality can also be improved by medical record quality control, because the medical record quality control not only can find and correct the problems in writing, but also can find the defects of doctors in the diagnosis process by analyzing medical records, and provide targeted guidance and training, thereby improving the medical level and the working efficiency of the doctors. Meanwhile, the standardized medical record writing is also beneficial to the communication and cooperation among doctors, so that patients can be treated more comprehensively and systematically. And moreover, the medical record quality control can also ensure the personal safety of a patient, and the standard medical record writing can avoid medical accidents caused by diagnosis and treatment errors and ensure the life safety and health rights of the patient. Meanwhile, the medical record quality control can also find potential medical hidden dangers and risks, measures are taken in time to prevent the medical hidden dangers and the risks, and therefore the occurrence rate of medical risks and patient safety accidents is reduced. Finally, medical record quality control is also beneficial to clinical scientific research and medical education, and standardized medical record writing can provide more accurate and complete data support for clinical scientific research, so that the progress and development of medical research are promoted. Meanwhile, through quality control and analysis of medical records, experience and deficiency of doctors in clinical practice can be found, beneficial materials and cases are provided for medical education, and medical education quality and level are improved.
Medical history content quality control refers to quality control of medical history content, and comprises quality control of a medical record first page, an admission record, an discharge record, a course of a disease, an doctor's advice, a consultation record and the like. The main task of medical record content control is to ensure the integrity, accuracy and standardization of medical record content so as to provide reliable diagnosis and treatment basis for doctors. Specific requirements for medical record content quality control include checking the logics and integrity of the home page, such as checking whether the data is full, whether diagnosis and operation are correct, whether omission exists, etc. Meanwhile, the logic of the main complaints and the diagnosis in the admission records are required to be checked, so that the rationality and the accuracy of the admission records are ensured. In addition, during the course of the disease, all medications and examinations need to be described in order for the physician to understand the course of treatment and changes in the condition of the patient. In order to improve the medical record writing quality, advanced departments such as a nursing part, a medical part and the like can take various measures to control the medical record content quality, including making and updating medical record writing standards, developing medical record writing training and assessment, regularly conducting spot check and evaluation on medical records, timely feeding back problems, improving advice and the like.
At present, the existing medical record content quality control method is mainly realized based on a named entity recognition and rule engine mode, and specifically takes a content quality control rule of 'main complaints and admission record diagnosis are not obviously associated' as an example. For an electronic medical record of an admission record, a named entity recognition model is firstly used for recognizing 'main complaints' and 'preliminary diagnosis' entities in the admission record, and then a rule engine is used for enumerating the corresponding relations between the main complaints and the preliminary diagnosis so as to control the quality of the medical record. But if it is not dependent on an exhaustive quality control rule, such as timeliness (admission records must be completed within 24 hours after admission), the method can better complete medical record quality control. However, if the quality control rule is exhaustive, the method needs to have no obvious correlation between the main complaint and the admission record diagnosis, but the number of diagnoses is often thousands of, and if the diagnosis is realized by depending on a rule engine, the quality control effect is difficult to obtain.
Therefore, in the existing medical record content quality control method, the rule engines are more relied on in the quality control rules which need to be exhausted, and the quality control effect is difficult to ensure when the diagnosis quantity is large.
Disclosure of Invention
Based on the above, it is necessary to provide a medical record inclusion quality control method, device, electronic equipment and storage medium, which can complete medical record inclusion quality control rules requiring exhaustion by using the context learning capability of a large language model and can guarantee quality control effect when the number of diagnoses is large.
The invention provides a medical record content quality control method, which comprises the following steps:
Acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
Taking a plurality of different first question texts as input of a large language model, and taking the first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training;
And receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
In one embodiment, the acquiring medical text data and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text includes:
Acquiring a history medical record text, and preprocessing the history medical record text to obtain medical text data, wherein the preprocessing at least comprises data cleaning and data filtering, and the medical text data at least comprises a patient complaint text and a doctor diagnosis text;
Analyzing the medical text data to generate the first question text and a first judgment basis corresponding to the first answer text according to the patient complaint text and the doctor diagnosis text;
Wherein the doctor diagnosis text includes the first judgment basis.
In one embodiment, the inputting the plurality of different first question texts as a large language model, and outputting the first answer text corresponding to each first question text as an output of the large language model, so that the large language model performs contextual learning training, includes:
Based on the first question text and the first judgment basis corresponding to the first answer text, acquiring a plurality of different first question texts, first answer texts corresponding to each first question text and first judgment basis corresponding to each first answer text;
And converting a plurality of groups of different first question texts, first answer texts and first judgment bases into a plurality of first training samples.
In one embodiment, the inputting the plurality of different first question texts as a large language model, and outputting the first answer text corresponding to each first question text as an output of the large language model, so that the large language model performs context learning training, includes:
Based on the plurality of first training samples, model input data and model output data of the large language model are obtained, the model input data is the first question text, and the model output data is the first answer text and a first judgment basis;
Taking the first question text as input of the large language model, taking the first answer text and a first judgment basis as output of the large language model, and performing contextual learning training on the large language model;
wherein the model parameters remain unchanged during the contextual learning training of the large language model.
In one embodiment, receiving a second question text of a medical record to be quality controlled, and invoking the big language model after the contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, which includes:
Acquiring the medical record to be controlled, and preprocessing the medical record to be controlled to obtain medical text data to be controlled of the medical record to be controlled, wherein the medical text data to be controlled comprises a patient complaint text to be controlled and a doctor diagnosis text to be controlled;
and analyzing and disassembling the medical text data to be controlled so as to generate a second judgment basis of the second question text and the second answer text according to the complaint text of the patient to be controlled and the diagnosis text of the doctor to be controlled.
In one embodiment, receiving a second question text of a medical record to be quality controlled, and invoking the large language model after the contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, including:
taking the second problem text as input of the large language model after the contextual learning training to call the large language model after the contextual learning training to perform understanding analysis on the second problem text;
And outputting the second answer text and the second judgment basis corresponding to the second answer text according to the second judgment basis.
In one embodiment, the method further comprises:
Based on the second question text and the second judgment basis corresponding to the second answer text, acquiring a plurality of different second question texts, second answer texts corresponding to each second question text and the second judgment basis corresponding to each second answer text;
Converting a plurality of groups of different second question texts, second answer texts and second judgment bases into a plurality of second training samples;
And when the large language model after the contextual learning training carries out understanding analysis on the second problem text, calling the large language model after the contextual learning training to carry out contextual learning training according to the second training sample.
The invention also provides a medical record content quality control device, which comprises:
the data analysis module is used for acquiring medical text data and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
The model training module is used for taking a plurality of different first question texts as input of a large language model, and taking the first answer text corresponding to each first question text as output of the large language model so as to enable the large language model to perform contextual learning training;
And the medical record quality control module is used for receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
The invention also provides an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the medical record content quality control method according to any one of the above when executing the computer program.
The invention also provides a computer storage medium storing a computer program which when executed by a processor realizes the medical record content quality control method according to any one of the above.
The invention also provides a computer program product, which comprises a computer program, wherein the computer program realizes the medical record content quality control method according to any one of the above when being executed by a processor.
According to the medical record content quality control method, the medical record content quality control device, the electronic equipment and the storage medium, medical text data are acquired, and the medical text data are analyzed, so that the medical text data are converted into corresponding question texts and answer texts corresponding to the question texts. And then, taking a plurality of different question texts as the input of the large language model, and taking the answer text corresponding to each question text as the output of the large language model, so that the large language model performs contextual learning training. And finally, receiving the question text of the medical record to be controlled, and calling the large language model after the contextual learning training to process the question text to be controlled so as to output the answer text corresponding to the question text to be controlled. The method can complete medical record content quality control rules requiring exhaustion by using the context learning capability of the large language model, only needs to give training sample (few shot) examples to conduct context learning in an reasoning stage, does not need to finely tune the large language model, and can ensure quality control effect when the diagnosis quantity is large.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a medical record content quality control method according to the present invention;
FIG. 2 is a schematic diagram of an overall flow chart of medical record content quality control according to the medical record content quality control method in an embodiment of the present invention;
FIG. 3 is a second schematic flow chart of the medical record content quality control method according to the present invention;
FIG. 4 is a third schematic flow chart of the medical record content quality control method according to the present invention;
FIG. 5 is a flowchart illustrating a method for controlling content of medical records according to the present invention;
FIG. 6 is a flowchart illustrating a method for controlling content of medical records according to the present invention;
FIG. 7 is a flowchart illustrating a method for controlling content of medical records according to the present invention;
FIG. 8 is a flowchart of a medical record content quality control method according to the present invention;
FIG. 9 is a schematic structural diagram of the medical record content quality control device provided by the invention;
fig. 10 is an internal structural diagram of a computer device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The medical record inclusion quality control method, the medical record inclusion quality control device, the electronic equipment and the storage medium are described below with reference to fig. 1 to 10.
As shown in fig. 1, in one embodiment, a medical record content quality control method includes the following steps:
step S110, acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text.
Specifically, the server acquires medical text data, and analyzes and disassembles the acquired medical text data to convert the medical text data into corresponding question text and answer text corresponding to the question text, namely a first question text and a first answer text corresponding to the first question text.
The medical text data comprises a patient complaint text and a doctor diagnosis text, the first question text is in a question form generated according to the patient complaint text and the doctor diagnosis text, and the first answer text is in an answer form obtained according to the doctor diagnosis text aiming at the first question text.
Referring to fig. 2, in a specific embodiment, the medical record content quality control method provided by the present invention takes "main complaints and admission record diagnoses have no obvious correlation" content quality control as an example, and constructs a context learning template, which includes: in the example section, in the middle of two groups of "# # # # # # # # #", three examples are designed, namely [ Few-shot example 1], [ Few-shot example 2 ], and [ Few-shot example 3 ], and each example is divided into a question and an answer. Problems: inquiring whether the main complaint and the admission main diagnosis are relevant or not, and answering: whether the association exists or not is answered first, and the specific reasons are analyzed. Wherein, [ Few-shot example 1], [ Few-shot example 2 ], and [ Few-shot example 3 ] are shown as follows:
###############
Few-shot example 1
Problem 1:
Complaints: fever for 7 days and cough for 5 days
Whether the complaint in the admission record is associated with the primary diagnosis of "mycoplasma pneumoniae"?
Answer 1: there is an association.
According to the following:
1. Symptoms of mycoplasma pneumoniae are: fever and cough
2. The complaint describes fever and cough, and refers to manifestations in mycoplasma pneumoniae
Therefore, the complaint "fever for 7 days and cough for 5 days" is associated with the main diagnosis of "mycoplasma pneumoniae pneumonia".
Few-shot example 2
Problem 2:
complaints: pain of left lower limb caused by injury and limited movement for 2 weeks
Whether or not the complaint in the admission record is associated with the primary diagnosis of "mental retardation?
Answer 2: no association.
According to the following:
1. Symptoms of mental retardation are: body weakness, reduced exercise amount, retarded reaction, stiffness of limbs, etc
2. Complaints describe swelling and pain of the left lower limb, restricted movement, inconsistent with mental retardation
Therefore, the main complaint of "swelling and pain of left lower limb due to fall injury, limited activity for 2 weeks" is not associated with the main diagnosis of "mental retardation".
Few-shot example 3
Problem 3:
Complaints: vomiting 1 day
Whether or not the complaint in the admission record is associated with the primary diagnosis of "allergic purpura"?
Answer 3: there is an association.
According to the following:
1. allergic purpura is usually manifested as: skin purpura, joint swelling, abdominal pain, emesis, etc
2. Vomiting is described in the complaint, referring to manifestations in allergic purpura and therefore, the complaint "vomiting for 1 day" is associated with the primary diagnosis of "allergic purpura".
###############
Step S120, a plurality of different first question texts are used as input of a large language model, and a first answer text corresponding to each first question text is used as output of the large language model, so that the large language model performs contextual learning training.
Specifically, the server takes a plurality of different first question texts obtained from the medical text data as input of a large language model, and takes a first answer text corresponding to each large question text as output of the large language model, so that the large language model performs contextual learning training. In the process of the contextual training, the generation of the first answer text is based on the generation basis of the first answer text, so that the large language model can understand and analyze the generation basis of the first answer text in the context training.
Step S130, receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
Specifically, the server receives a question text of the medical record to be quality controlled, namely a second question text, and invokes the context trained large language model to process the second question text, so as to output an answer text corresponding to the second question text, namely the second answer text, according to the patient complaint text and the doctor diagnosis text in the medical record to be quality controlled, and meanwhile, the user can also check the basis obtained by the second answer text.
Referring to fig. 2, in a specific embodiment, the medical record content quality control method provided by the invention needs a problem part of medical record with quality control, gives a main complaint and a main diagnosis of admission, and enables a large model to answer whether the answer is relevant. Taking the open source large language model base Qwen-14B-Chat as an example, the large language model does not answer well without the context learning example. But training the large language model by the context learning method, the large language model can have the capability of controlling the content quality of the new medical record, for example, the large language model answers as follows:
Problems:
complaints: abnormal urine test for 5 months
Is the complaint in the admission record associated with the primary diagnosis of "chemotherapy"? If the association exists, answering the basis of the association and the association; if the data is not associated, answering the basis of 'no association' and no association;
Answer: no association.
According to the following: chemotherapy is a therapeutic method for killing cancer cells by using drugs, and is generally used for treating cancer. While abnormal urine tests may be due to a variety of causes, such as urinary system infections, kidney stones, diabetes, etc., are not necessarily associated with chemotherapy. Thus, the complaint "abnormal urine detection for 5 months" was not associated with the main diagnosis of "chemotherapy".
At this time, the medical record to be quality-controlled can be judged to be a defective medical record.
In this embodiment, in addition to the above-mentioned "no obvious correlation between complaints and admission records" the following content quality control rule description is given:
under the quality control rule of judging whether the operation history has the operation which can cause scars, if the operation is dependent on a rule engine, all operations are needed to be exhausted, the operations are difficult to complete, and the inclusion quality control can be completed by using the context learning capability of a large model. For example:
#################
Example 1:
history of surgery: deny history of surgical trauma
Asking for: judging whether the surgical history can cause surgical scars or surgical incisions, responding or not, and analyzing.
Answering: it will not be denied because of a history of surgical trauma.
Example 2:
history of surgery: with history of surgical trauma, neonatal intracranial hemorrhage was asked in 2021.10.18 China's general anesthesia downstream craniotomy subdural hematoma, and intracerebral hematoma removal surgery: judging whether the surgical history can cause surgical scars or surgical incisions, responding or not, and analyzing.
Answering: there is a description in the history of the operation that "general anesthesia down to craniotomy subdural hematoma and intracerebral hematoma removal in 2021.10.18. This causes surgical scars.
#################
Now please answer:
History of surgery: with history of surgical trauma, the new born intracranial hemorrhage is treated by general anesthesia descending craniotomy subdural hematoma and intracerebral hematoma removal in 2021.10.18 China
Asking for: judging whether the surgical history can cause surgical scars or surgical incisions, responding or not, and analyzing.
Answering: * Output of large language model
In addition, under the quality control rule that the "admission diagnosis has no corresponding diagnosis basis" in the admission record, the rule engine is used for exhaustive diagnosis, and the context learning capability of the large model can be used for completing the content quality control, which is not repeated here.
In this embodiment, the problem that the medical record content quality control rule needs to be exhausted can be solved, the experimental result shows that part of the medical record content quality control rule needs to be exhausted, the defect medical record is found out to be less than 0.01% by using the prior art, the defect medical record accounts for about 0.2% -0.5% by using the contextual learning capability of the large language model, and the average accuracy rate of the found defect medical record can reach 80% through the examination of an expert, so that the range of medical record content quality control is greatly expanded, and the accuracy rate of medical record content quality control is improved.
According to the medical record content quality control method, medical text data are acquired and analyzed, so that the medical text data are converted into corresponding question texts and answer texts corresponding to the question texts. And then, taking a plurality of different question texts as the input of the large language model, and taking the answer text corresponding to each question text as the output of the large language model, so that the large language model performs contextual learning training. And finally, receiving the question text of the medical record to be controlled, and calling the large language model after the contextual learning training to process the question text to be controlled so as to output the answer text corresponding to the question text to be controlled. The method can complete the medical record content quality control rule which needs to be exhausted by using the context learning capability of the large language model, only needs to give a training sample (Few-shot) example to conduct context learning in an reasoning stage, does not need to finely tune the large language model, and can ensure the quality control effect when the diagnosis quantity is large.
As shown in fig. 3, in one embodiment, the medical record content quality control method provided by the present invention acquires medical text data, and analyzes the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text, and specifically includes the following steps:
Step S112, acquiring a history medical record text, and preprocessing the history medical record text to obtain medical text data, wherein the preprocessing at least comprises data cleaning and data filtering, and the medical text data at least comprises a patient complaint text and a doctor diagnosis text.
Specifically, the server acquires a history medical record text, and performs cleaning and filtering pretreatment on the history medical record text to obtain medical text data at least comprising a patient complaint text and a doctor diagnosis text.
Step S114, analyzing the medical text data to generate a first question text and a first judgment basis corresponding to the first answer text according to the patient complaint text and the doctor diagnosis text.
Specifically, the server analyzes and disassembles the medical text data obtained in step S112, so as to generate a corresponding question text, an answer text corresponding to the question text, and a judgment basis corresponding to the answer text, i.e., a first question text, a first answer text, and a first judgment basis corresponding to the first answer text, according to the patient complaint text and the doctor diagnosis text.
As shown in fig. 4, in one embodiment, the medical record content quality control method provided by the present invention uses a plurality of different first question texts as input of a large language model, and uses a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training, and includes the following steps:
Step S410, based on the first question text and the first judgment basis corresponding to the first answer text, obtains a plurality of different first question texts, the first answer text corresponding to each first question text, and the first judgment basis corresponding to each first answer text.
Specifically, the server obtains a plurality of different first question texts, first answer texts corresponding to each first question text and first judgment bases corresponding to each first answer text based on the first question texts and first judgment bases corresponding to the first question texts.
In step S420, a plurality of sets of different first question texts, first answer texts and first judgment bases are converted into a plurality of first training samples.
Specifically, the server converts multiple groups of different and mutually corresponding first question texts, first answer texts and first judgment bases into multiple training samples (Few-shot), namely first training samples.
As shown in fig. 5, in one embodiment, the medical record content quality control method provided by the present invention uses a plurality of different first question texts as input of a large language model, and uses a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs context learning training, and specifically includes the following steps:
Step S510, based on a plurality of first training samples, model input data and model output data of the large language model are obtained, the model input data is a first question text, and the model output data is a first answer text and a first judgment basis.
Specifically, the server obtains model input data and model output data of the large language model based on a plurality of first training samples obtained previously, wherein the model input data is a first question text, and the model output data is a first answer text and a first judgment basis.
Step S520, the first question text is used as the input of the large language model, and the first answer text and the first judgment basis are used as the output of the large language model to perform the context learning training on the large language model.
Specifically, the server takes the first question text as input of the large language model, and takes the first answer text and the first judgment basis as output of the large language model, so that the large language model outputs the large answer text based on analysis of the first question text and the first judgment basis, and carries out contextual learning training on the large language model.
As shown in fig. 6, in one embodiment, the medical record content quality control method provided by the present invention receives a second question text of a medical record to be quality controlled, and invokes a big language model after contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, and the method includes the following steps:
Step S610, obtaining medical records to be controlled, and preprocessing the medical records to be controlled to obtain medical text data to be controlled of the medical records to be controlled, wherein the medical text data to be controlled comprises a patient complaint text to be controlled and a doctor diagnosis text to be controlled.
Specifically, the server acquires medical records to be controlled, and performs pretreatment of cleaning and filtering on the medical records to be controlled to obtain medical text data of the medical records to be controlled, namely medical text data to be controlled, wherein the medical text data to be controlled comprises a patient complaint text to be controlled and a doctor diagnosis text to be controlled.
Step S620, analyzing and disassembling the medical text data to be controlled so as to generate a second question text and a second judgment basis of a second answer text according to the main complaint text of the patient to be controlled and the diagnosis text of the doctor to be controlled.
Specifically, the server analyzes and disassembles the medical text data to be controlled obtained in step S610, so as to generate a corresponding question form according to the patient complaint text to be controlled and the doctor diagnosis text to be controlled, and obtain a second question text, a second answer text corresponding to the second question text, and a second judgment basis corresponding to the second answer text.
As shown in fig. 7, in one embodiment, the medical record content quality control method provided by the present invention receives a second question text of a medical record to be quality controlled, and invokes a big language model after contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, and specifically includes the following steps:
step S132, taking the second question text as the input of the large language model after the contextual learning training to call the large language model after the contextual learning training to perform understanding analysis on the second question text.
Specifically, the server takes the second problem text as input of the large language model after the contextual learning training to call the large language model after the contextual learning training to understand and analyze the second problem text.
Step S134, outputting the second answer text and the second judgment basis corresponding to the second answer text according to the second judgment basis.
Specifically, the server outputs a final second answer text according to the second judgment basis on the basis of understanding and analyzing the second question text by calling the context learning trained large language model, and then outputs the second judgment basis corresponding to the second answer text.
As shown in fig. 8, in one embodiment, the medical record content quality control method provided by the present invention further includes the following steps:
Step 810, based on the second question text and the second judgment basis corresponding to the second answer text, obtaining a plurality of different second question texts, the second answer text corresponding to each second question text, and the second judgment basis corresponding to each second answer text.
Specifically, the server obtains a plurality of groups of different and mutually corresponding second question texts, second answer texts corresponding to each second question text and second judgment bases corresponding to each second answer text based on the second question texts and second judgment bases of the second question texts in the answer reasoning process.
Step S820, converting the plurality of sets of different second question texts, second answer texts and second judgment basis into a plurality of second training samples.
Specifically, the server converts multiple groups of different second question texts, second answer texts and second judgment bases which correspond to each other into multiple training samples (Few-shot), namely second training samples.
In step S830, when the large language model after the contextual training carries out understanding analysis on the second problem text, the large language model after the contextual training is called to carry out contextual training according to the second training sample.
Specifically, in the process that the server carries out understanding analysis on the second problem text by the large language model after the contextual learning training, the large language model after the contextual learning training is repeatedly called to carry out contextual learning training according to a second training sample formed by the medical record to be quality controlled later, so that in the process of quality control of more and more medical records to be quality controlled, the large language model is repeatedly subjected to contextual learning training, and the quality control precision of the large language model is improved.
The medical record content quality control device provided by the invention is described below, and the medical record content quality control device described below and the medical record content quality control method described above can be correspondingly referred to each other.
As shown in FIG. 9, in one embodiment, a medical record content quality control apparatus includes a data analysis module 910, a model training module 920, and a medical record quality control module 930.
The data analysis module 910 is configured to obtain medical text data, and analyze the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text.
The model training module 920 is configured to take a plurality of different first question texts as input of the large language model, and take a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training.
The medical record quality control module 930 is configured to receive a second question text of the medical record to be quality controlled, and invoke the context learning trained large language model to process the second question text, so as to output a second answer text corresponding to the second question text.
In this embodiment, the medical record content quality control device provided by the invention, the data analysis module is specifically configured to:
And acquiring the history medical record text, and preprocessing the history medical record text to obtain medical text data, wherein the preprocessing at least comprises data cleaning and data filtering, and the medical text data at least comprises a patient complaint text and a doctor diagnosis text.
And analyzing the medical text data to generate a first question text and a first judgment basis corresponding to the first answer text according to the patient complaint text and the doctor diagnosis text.
Wherein, doctor diagnosis text comprises a first judgment basis.
In this embodiment, the medical record content quality control device provided by the present invention further includes a training sample obtaining module, configured to:
Based on the first question text and the first judgment basis corresponding to the first answer text, a plurality of different first question texts, the first answer text corresponding to each first question text and the first judgment basis corresponding to each first answer text are obtained.
And converting a plurality of groups of different first question texts, first answer texts and first judgment bases into a plurality of first training samples.
In this embodiment, the medical record content quality control device provided by the invention, the model training module is specifically configured to:
Based on a plurality of first training samples, model input data and model output data of a large language model are obtained, the model input data is a first question text, and the model output data is a first answer text and a first judgment basis.
And taking the first question text as the input of the large language model, taking the first answer text and the first judgment basis as the output of the large language model, and performing context learning training on the large language model.
Wherein the model parameters remain unchanged during the training of the contextual learning of the large language model.
In this embodiment, the medical record content quality control device provided by the invention further includes a medical record preprocessing module for quality control, which is used for:
And acquiring medical records to be controlled, and preprocessing the medical records to be controlled to obtain medical text data to be controlled of the medical records to be controlled, wherein the medical text data to be controlled comprises a patient complaint text to be controlled and a doctor diagnosis text to be controlled.
Analyzing and disassembling the medical text data to be controlled so as to generate a second question text and a second judgment basis of a second answer text according to the complaint text of the patient to be controlled and the diagnosis text of the doctor to be controlled.
In this embodiment, the medical record content quality control device provided by the invention, the medical record quality control module is specifically configured to:
And taking the second question text as the input of the large language model after the contextual learning training to call the large language model after the contextual learning training to perform understanding analysis on the second question text.
And outputting the second answer text and the second judgment basis corresponding to the second answer text according to the second judgment basis.
In this embodiment, the medical record content quality control device provided by the invention further includes a model training sub-module, which is specifically configured to:
Based on the second question text and the second judgment basis corresponding to the second answer text, a plurality of different second question texts, the second answer text corresponding to each second question text and the second judgment basis corresponding to each second answer text are obtained.
And converting a plurality of groups of different second question texts, second answer texts and second judgment bases into a plurality of second training samples.
And when the large language model after the contextual training carries out understanding analysis on the second problem text, calling the large language model after the contextual training to carry out contextual training according to the second training sample.
Fig. 10 illustrates a physical structure diagram of an electronic device, which may be an intelligent terminal, and an internal structure diagram thereof may be as shown in fig. 10. The electronic device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the electronic device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements a medical record content quality control method, the method comprising:
Acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
Taking a plurality of different first question texts as input of a large language model, and taking a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training;
and receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
It will be appreciated by those skilled in the art that the structure shown in fig. 10 is merely a block diagram of a portion of the structure associated with the present inventive arrangements and is not limiting of the electronic device to which the present inventive arrangements are applied, and that a particular electronic device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In another aspect, the present invention further provides a computer storage medium storing a computer program, where the computer program when executed by a processor implements a medical record content quality control method, the method includes:
Acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
Taking a plurality of different first question texts as input of a large language model, and taking a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training;
and receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
In yet another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. A processor of an electronic device reads the computer instructions from a computer readable storage medium, and when the processor executes the computer instructions, the processor implements a medical record content quality control method, the method comprising:
Acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
Taking a plurality of different first question texts as input of a large language model, and taking a first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training;
and receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory.
By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the invention and are described in detail herein without thereby limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of protection of the present invention is to be determined by the appended claims.
Claims (10)
1. A medical record inclusion quality control method, which is characterized by comprising the following steps:
Acquiring medical text data, and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
Taking a plurality of different first question texts as input of a large language model, and taking the first answer text corresponding to each first question text as output of the large language model, so that the large language model performs contextual learning training;
And receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
2. The medical record content quality control method according to claim 1, wherein the acquiring medical text data and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text comprises:
Acquiring a history medical record text, and preprocessing the history medical record text to obtain medical text data, wherein the preprocessing at least comprises data cleaning and data filtering, and the medical text data at least comprises a patient complaint text and a doctor diagnosis text;
Analyzing the medical text data to generate the first question text and a first judgment basis corresponding to the first answer text according to the patient complaint text and the doctor diagnosis text;
Wherein the doctor diagnosis text includes the first judgment basis.
3. The medical record content quality control method according to claim 2, wherein the step of taking a plurality of different first question texts as inputs of a large language model and taking the first answer text corresponding to each first question text as an output of the large language model, so that the large language model performs contextual learning training, comprises:
Based on the first question text and the first judgment basis corresponding to the first answer text, acquiring a plurality of different first question texts, first answer texts corresponding to each first question text and first judgment basis corresponding to each first answer text;
And converting a plurality of groups of different first question texts, first answer texts and first judgment bases into a plurality of first training samples.
4. The medical record content quality control method according to claim 3, wherein the step of using a plurality of different first question texts as inputs of a large language model and using the first answer text corresponding to each of the first question texts as outputs of the large language model to enable the large language model to perform a context learning training comprises:
Based on the plurality of first training samples, model input data and model output data of the large language model are obtained, the model input data is the first question text, and the model output data is the first answer text and a first judgment basis;
Taking the first question text as input of the large language model, taking the first answer text and a first judgment basis as output of the large language model, and performing contextual learning training on the large language model;
wherein the model parameters remain unchanged during the contextual learning training of the large language model.
5. The medical record content quality control method according to claim 1, wherein receiving a second question text of a medical record to be quality controlled, and invoking the large language model after the contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, includes:
Acquiring the medical record to be controlled, and preprocessing the medical record to be controlled to obtain medical text data to be controlled of the medical record to be controlled, wherein the medical text data to be controlled comprises a patient complaint text to be controlled and a doctor diagnosis text to be controlled;
and analyzing and disassembling the medical text data to be controlled so as to generate a second judgment basis of the second question text and the second answer text according to the complaint text of the patient to be controlled and the diagnosis text of the doctor to be controlled.
6. The medical record content quality control method according to claim 5, wherein receiving a second question text of a medical record to be quality controlled, and invoking the large language model after the contextual learning training to process the second question text, so as to output a second answer text corresponding to the second question text, comprises:
taking the second problem text as input of the large language model after the contextual learning training to call the large language model after the contextual learning training to perform understanding analysis on the second problem text;
And outputting the second answer text and the second judgment basis corresponding to the second answer text according to the second judgment basis.
7. The medical record content quality control method according to claim 6, further comprising:
Based on the second question text and the second judgment basis corresponding to the second answer text, acquiring a plurality of different second question texts, second answer texts corresponding to each second question text and the second judgment basis corresponding to each second answer text;
Converting a plurality of groups of different second question texts, second answer texts and second judgment bases into a plurality of second training samples;
And when the large language model after the contextual learning training carries out understanding analysis on the second problem text, calling the large language model after the contextual learning training to carry out contextual learning training according to the second training sample.
8. A medical record content quality control device, characterized in that the device comprises:
the data analysis module is used for acquiring medical text data and analyzing the medical text data to convert the medical text data into a first question text and a first answer text corresponding to the first question text;
The model training module is used for taking a plurality of different first question texts as input of a large language model, and taking the first answer text corresponding to each first question text as output of the large language model so as to enable the large language model to perform contextual learning training;
And the medical record quality control module is used for receiving a second question text of the medical record to be quality controlled, and calling the large language model after the contextual learning training to process the second question text so as to output a second answer text corresponding to the second question text.
9. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 7 when the computer program is executed.
10. A computer storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410406684.9A CN117995347B (en) | 2024-04-07 | 2024-04-07 | Medical record content quality control method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410406684.9A CN117995347B (en) | 2024-04-07 | 2024-04-07 | Medical record content quality control method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117995347A true CN117995347A (en) | 2024-05-07 |
| CN117995347B CN117995347B (en) | 2024-06-21 |
Family
ID=90887758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410406684.9A Active CN117995347B (en) | 2024-04-07 | 2024-04-07 | Medical record content quality control method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117995347B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108172301A (en) * | 2018-01-31 | 2018-06-15 | 中国科学院软件研究所 | A kind of mosquito matchmaker's epidemic Forecasting Methodology and system based on gradient boosted tree |
| WO2021139424A1 (en) * | 2020-05-14 | 2021-07-15 | 平安科技（深圳）有限公司 | Text content quality evaluation method, apparatus and device, and storage medium |
| KR20210090576A (en) * | 2020-11-30 | 2021-07-20 | 베이징 바이두 넷컴 사이언스 앤 테크놀로지 코., 엘티디. | A method, an apparatus, an electronic device, a storage medium and a program for controlling quality |
| WO2021188509A1 (en) * | 2020-03-17 | 2021-09-23 | Decoded Health, Inc. | Machine-assisted medical patient interaction, diagnosis, and treatment |
| CN114664431A (en) * | 2022-03-15 | 2022-06-24 | 平安科技（深圳）有限公司 | Artificial intelligence assisted question and answer method, device, equipment and medium |
| CN115631852A (en) * | 2022-11-02 | 2023-01-20 | 北京大学重庆大数据研究院 | Certificate type recommendation method and device, electronic equipment and nonvolatile storage medium |
| CN116092699A (en) * | 2021-11-05 | 2023-05-09 | 上海仰和华健人工智能科技有限公司 | Cancer question-answer interaction method based on pre-training model |
| CN117454989A (en) * | 2023-11-14 | 2024-01-26 | 生命奇点(北京)科技有限公司 | System for updating electronic medical record question-answer model based on parameter adjustment |
| CN117577254A (en) * | 2023-11-17 | 2024-02-20 | 上海交通大学医学院附属瑞金医院 | Method and system for constructing language model in medical field and structuring text of electronic medical record |
| WO2024042348A1 (en) * | 2022-08-24 | 2024-02-29 | Evyd科技有限公司 | English medical text structurization method, apparatus, medium and electronic device |
| CN117708306A (en) * | 2024-02-06 | 2024-03-15 | 神州医疗科技股份有限公司 | Medical question-answering architecture generation method and system based on layered question-answering structure |
-
2024
- 2024-04-07 CN CN202410406684.9A patent/CN117995347B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108172301A (en) * | 2018-01-31 | 2018-06-15 | 中国科学院软件研究所 | A kind of mosquito matchmaker's epidemic Forecasting Methodology and system based on gradient boosted tree |
| WO2021188509A1 (en) * | 2020-03-17 | 2021-09-23 | Decoded Health, Inc. | Machine-assisted medical patient interaction, diagnosis, and treatment |
| WO2021139424A1 (en) * | 2020-05-14 | 2021-07-15 | 平安科技（深圳）有限公司 | Text content quality evaluation method, apparatus and device, and storage medium |
| KR20210090576A (en) * | 2020-11-30 | 2021-07-20 | 베이징 바이두 넷컴 사이언스 앤 테크놀로지 코., 엘티디. | A method, an apparatus, an electronic device, a storage medium and a program for controlling quality |
| CN116092699A (en) * | 2021-11-05 | 2023-05-09 | 上海仰和华健人工智能科技有限公司 | Cancer question-answer interaction method based on pre-training model |
| CN114664431A (en) * | 2022-03-15 | 2022-06-24 | 平安科技（深圳）有限公司 | Artificial intelligence assisted question and answer method, device, equipment and medium |
| WO2024042348A1 (en) * | 2022-08-24 | 2024-02-29 | Evyd科技有限公司 | English medical text structurization method, apparatus, medium and electronic device |
| CN115631852A (en) * | 2022-11-02 | 2023-01-20 | 北京大学重庆大数据研究院 | Certificate type recommendation method and device, electronic equipment and nonvolatile storage medium |
| CN117454989A (en) * | 2023-11-14 | 2024-01-26 | 生命奇点(北京)科技有限公司 | System for updating electronic medical record question-answer model based on parameter adjustment |
| CN117577254A (en) * | 2023-11-17 | 2024-02-20 | 上海交通大学医学院附属瑞金医院 | Method and system for constructing language model in medical field and structuring text of electronic medical record |
| CN117708306A (en) * | 2024-02-06 | 2024-03-15 | 神州医疗科技股份有限公司 | Medical question-answering architecture generation method and system based on layered question-answering structure |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117995347B (en) | 2024-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yorke et al. | The association between emotional and behavioral problems in children with autism spectrum disorder and psychological distress in their parents: A systematic review and meta-analysis | |
| Park et al. | Comparison of methods to reduce bias from clinical prediction models of postpartum depression | |
| Bradbury et al. | Attributions and behavior in functional and dysfunctional marriages. | |
| KR102237449B1 (en) | Method, server and program of learning a patient diagnosis | |
| Wen et al. | Associations of assisted reproductive technology and twin pregnancy with risk of congenital heart defects | |
| CN109887596A (en) | Chronic obstructive disease of lung diagnostic method, device and the computer equipment of knowledge based map | |
| Hoyos et al. | Measures of patient dissatisfaction with health care in polycystic ovary syndrome: retrospective analysis | |
| KR102332740B1 (en) | Method, apparatus and system for diagnosing patient's symptoms using deep learning | |
| Åkerblom et al. | A validation and generality study of the committed action questionnaire in a Swedish sample with chronic pain | |
| Gudu et al. | Development of a medical expert system as an expertknowledge sharing tool on diagnosis and treatment of hypertension in pregnancy | |
| CN113611433A (en) | Auxiliary system and method for inquiry | |
| CN113393945A (en) | Clinical drug allergy management method, auxiliary device and system | |
| CN117995340B (en) | Intelligent recruitment method and device for clinical trial based on large model | |
| Nettelbeck et al. | Speed of information processing and cognition | |
| CN114628034A (en) | Alzheimer's disease assessment method, system, device and storage medium | |
| CN116612879B (en) | Diagnostic result prediction method, diagnostic result prediction device, electronic equipment and storage medium | |
| CN117995347B (en) | Medical record content quality control method and device, electronic equipment and storage medium | |
| Saltapidas et al. | The influence of cultural background on experiences and beliefs about traumatic brain injury and their association with outcome | |
| Sahoo et al. | Scientific reproducibility in biomedical research: provenance metadata ontology for semantic annotation of study description | |
| Kopal et al. | Endorsing complexity through diversity: computational psychiatry meets big data analytics | |
| Windriyani et al. | Expert system for detecting mental disorder with Forward Chaining method | |
| Gunn et al. | Narratives of gestational diabetes provide a lens to tailor postpartum prevention and monitoring counseling | |
| Liao et al. | A machine learning‐based risk scoring system for infertility considering different age groups | |
| Hashmani et al. | Distribution and correlation of ocular surface disease index scores in a non-clinical population: the Karachi ocular surface disease study | |
| Luo et al. | Towards Accurate and Clinically Meaningful Summarization of Electronic Health Record Notes: A Guided Approach |
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 |