CN113611401A - Perioperative blood management system and method - Google Patents
Perioperative blood management system and method Download PDFInfo
- Publication number
- CN113611401A CN113611401A CN202110986673.9A CN202110986673A CN113611401A CN 113611401 A CN113611401 A CN 113611401A CN 202110986673 A CN202110986673 A CN 202110986673A CN 113611401 A CN113611401 A CN 113611401A
- Authority
- CN
- China
- Prior art keywords
- blood
- data
- patient
- module
- artificial intelligence
- 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
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
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- General Business, Economics & Management (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The embodiment of the invention discloses a perioperative blood management system and a perioperative blood management method, which comprise an information acquisition module, an information processing module, an artificial intelligence calculation module and a clinical support result display module, wherein the information acquisition module acquires diagnosis and treatment data of a patient in real time; the information processing module extracts important blood evaluation features based on diagnosis and treatment data of a patient and converts the important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning; the artificial intelligence calculation module constructs a classification model based on the characteristics recognizable by the artificial intelligence engine, so as to realize the assessment of anemia before an operation, the prediction of perioperative blood transfusion and the prediction of risks of adverse events of postoperative blood transfusion, and obtain a blood use assessment result; and the clinical support result display module integrates the blood use evaluation result and carries out early warning and recommendation according to the blood use evaluation result. Therefore, the invention realizes multi-link coverage and dynamic monitoring of perioperative blood management, and reduces the transfusion risk to the maximum extent.
Description
Technical Field
The invention relates to the technical field of applying artificial intelligence in the medical field, in particular to a perioperative blood management system and a perioperative blood management method.
Background
Blood transfusion is an irreplaceable clinical treatment for life saving, and a large amount of blood products are used clinically every day, but the use thereof faces huge challenges. On the one hand, blood product resources are limited, while at the same time the total amount of transfusions continues to increase, of which 70% are used in surgical patients. The supply of blood is difficult to meet the increasing demand and blood can be preserved for only a short time. On the other hand, there is a lot of evidence that perioperative allotransfusion increases morbidity and mortality in surgical patients, reducing long-term survival, with higher expenditure and more resource usage. The effective, safe and scientific use of blood, the maximum avoidance of blood waste and the guarantee of patient safety have become one of the focuses of medical industry attention.
The blood management of patients is based on the concept of syndrome-based medicine, and by taking the patients as the center and applying a multidisciplinary combined method, the blood loss is reduced to the maximum extent by preventing and treating anemia and improving the blood coagulation function, so that the allogenic blood transfusion is reduced or avoided, and the clinical outcome of the patients is improved. Administration of blood management to a patient can reduce transfusion-related risks and improve patient prognosis.
However, the current perioperative blood management mode is still not standard enough, and perioperative blood is guided by clinical experience to a large extent, so that the perioperative blood infusion rate level is higher. Therefore, not only is a serious blood waste condition caused, but also the safety of perioperative patients is threatened. Therefore, blood management aids have come into play.
The blood management auxiliary Tool adopted in the existing practice is mainly blood Transfusion Risk Scoring, including Scoring such as a Transfusion Risk Understanding Scoring Tool (TRUST), or a Transfusion Risk and Clinical Knowledge (TRACK), and the basic variables are taken into a Scoring system to determine a prediction threshold value, so that the Transfusion Risk is divided, and the use is simple and convenient. However, although the adopted blood management aid can predict the transfusion risk and identify high-risk patients to facilitate risk adjustment and blood management planning, it mainly aims at the transfusion risk, and lacks integration of links such as anemia assessment before operation, blood transfusion prediction during operation, adverse event warning after operation, and the like.
Disclosure of Invention
In view of this, the embodiment of the present invention provides a perioperative blood management system, which can achieve multi-link coverage and dynamic monitoring of perioperative blood management, and reduce blood transfusion risk to the maximum extent.
The embodiment of the invention also provides a perioperative blood management method, which can realize multi-link coverage and dynamic monitoring of perioperative blood management and reduce blood transfusion risks to the maximum extent.
The invention is realized by the following steps:
a system for perioperative blood management, comprising: an information acquisition module, an information processing module, an artificial intelligence calculation module and a clinical support result display module, wherein,
the information acquisition module is used for acquiring diagnosis and treatment data of a patient in real time;
the information processing module is used for extracting important blood evaluation features based on diagnosis and treatment data of a patient and converting the important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning;
the artificial intelligence calculation module is used for constructing a classification model based on the characteristics recognizable by the artificial intelligence engine, realizing the assessment of anemia before an operation, the prediction of perioperative blood transfusion and the prediction of risks of adverse events of blood transfusion after the operation, and obtaining a blood use assessment result;
and the clinical support result display module is used for integrating the blood use evaluation results and carrying out early warning and recommendation according to the blood use evaluation results.
Preferably, the system further comprises: a dynamic monitoring module and an iterative optimization module, wherein,
the dynamic monitoring module is used for monitoring a system use log of a user, feeding back a blood use evaluation result and decision data of a blood use management process in real time, obtaining evaluation information and sending the abnormal information to the iterative optimization module when the evaluation information is abnormal;
and the iterative optimization module is used for triggering periodically or by received abnormal information, the artificial intelligence calculation module is used for training and updating the classification model, if the data expansion requirement exists, the information acquisition module is used for completing data increment, and then newly added data or knowledge is input into the artificial intelligence calculation module for training and updating the classification model, so that the system is continuously optimized.
Preferably, the information acquisition module is further configured to obtain diagnosis and treatment data of a patient in real time, and includes:
acquiring the identity of a patient; collecting a test record of a patient based on the identity of the patient through a Laboratory Information System (LIS); acquiring operation related data of a patient in an operation based on the identity of the patient through an operation anesthesia information system; acquiring the imaging examination data of the patient based on the identity of the patient through an image archiving and communication system PACS; acquiring the past medical history, the medical course record, the diagnosis and the medication record of a patient based on the identity of the patient through an electronic medical record system EMR and a hospital information system HIS; blood usage information is collected for the patient by the blood dispensing system based on the patient's identification.
Preferably, the information processing module is further configured to, according to the data range determined by the blood management literature search result, perform structuring processing on unstructured data within the blood management data range, and then perform preprocessing on all structured data to convert the structured data into a data form that can be recognized by the artificial intelligence calculation module;
the process of carrying out the structuring processing on the unstructured data comprises the following steps: converting printing characters in a paper file into an image formed by a black and white dot matrix by using an Optical Character Recognition (OCR) technology, and then recognizing characters in the image to convert the characters into text data; utilizing a natural language processing NLP technology, customizing a medical dictionary of diseases, vital signs and symptoms in advance, performing word segmentation processing on unstructured text data, extracting a theme of the text data after word segmentation, and further performing named entity recognition and relationship judgment; extracting sign values or judging whether diseases exist or not by using a regular expression, and realizing information conversion and structured representation of unstructured data;
the preprocessing is carried out on all the structured data to convert the structured data into a data form which can be identified by the artificial intelligence calculation module, and the data form is subjected to coding mapping on the structured data.
Preferably, the artificial intelligence computation module further comprises a recommendation model and a prediction model, wherein,
the recommendation model is used for analyzing the characteristics of the previous patient, calculating the similarity between the current patient and the previous patient, acquiring the previous patient most similar to the current patient, generating candidate items according to the medication or operation of the similar patients, and further improving the operation recommendation such as medication of the patient by combining a knowledge graph;
the prediction model is used for splitting diagnosis and treatment data of a previous patient into a training set and a test set during training, performing exhaustive search on hyperparameters of the model by using a grid search method to establish the model, and selecting an optimal hyperparameter combination and a threshold according to model performances on the test set; after training, transmitting the characteristics of the current patient into a model to obtain a prediction result, wherein the prediction result comprises the blood transfusion volume, the blood transfusion risk probability value and the adverse event risk probability value, and if the risk probability value is higher than a set threshold value, carrying out early warning prompt.
Preferably, the clinical support result display module is further configured to integrate the blood use evaluation result obtained by the artificial intelligence calculation module into a clinical decision support system CDSS after being artificially confirmed by the transfusion department, display a medication recommendation scheme for the anemia patient and a hand anesthesia recommendation operation during the operation, prompt perioperative blood preparation and blood inventory, and perform early warning on the transfusion and high risk patients with adverse events.
Preferably, the dynamic monitoring module further includes: a monitoring data collection unit, an evaluation index calculation unit and an early warning unit, wherein,
the monitoring data collection unit is used for collecting and storing the feedback result of the blood transfusion department confirmation step before result display, the collected clinical staff use log data and the patient blood vessel administration decision data in real time;
the evaluation index calculation unit is used for automatically calculating multi-dimensional evaluation indexes including but not limited to accuracy, usability, acceptance rate and the like in real time aiming at different personnel such as clinical personnel, transfusion departments and the like according to the monitoring data collected by the monitoring data collection unit;
and the early warning unit is used for drawing a statistical chart to display the index change in real time and carrying out early warning when any evaluation index is lower than a set threshold value according to the threshold value set by each index.
Preferably, the iterative optimization module further comprises a timing iteration unit and a specific condition triggering iteration unit, wherein,
the timing iteration unit is used for collecting new data generated in a period when the system is used for a certain period, expanding a database, updating and perfecting a knowledge map and performing dynamic optimization on the system;
and the specific condition triggering iteration unit is used for adjusting the artificial intelligence calculation module according to specific conditions when the evaluation index triggers early warning or clinical change requirements, and if the data expansion requirements exist, the information acquisition module completes data increment and then completes optimization.
A method of perioperative blood management, the method comprising:
acquiring diagnosis and treatment data of a patient in real time;
extracting important blood evaluation features based on diagnosis and treatment data of a patient, and cleaning and converting the important blood evaluation features into features recognizable by an artificial intelligence engine;
constructing a classification model based on the recognizable characteristics of an artificial intelligence engine, and realizing assessment of anemia before an operation, prediction of perioperative blood transfusion and prediction of risks of adverse events of blood transfusion after the operation to obtain a blood use assessment result;
and integrating blood evaluation results, and carrying out early warning and recommendation according to the evaluation results.
Preferably, the method further comprises: monitoring a system use log of a user in real time, feeding back a blood use evaluation result and decision data of a blood use management process, and obtaining evaluation information; when the evaluation information has abnormality, acquiring abnormality information; and triggering periodically or by received abnormal information, inputting newly added data or knowledge, training and updating the classification model, and continuously optimizing.
As can be seen from the above, the embodiment of the present invention provides a perioperative blood management system, which includes an information acquisition module, an information processing module, an artificial intelligence calculation module, and a clinical support result display module, wherein the information acquisition module acquires diagnosis and treatment data of a patient in real time; the information processing module extracts important blood evaluation features based on diagnosis and treatment data of a patient and converts the important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning; the artificial intelligence calculation module constructs a classification model based on the characteristics recognizable by the artificial intelligence engine, and obtains a blood use evaluation result after realizing the evaluation of anemia before an operation, the prediction of blood transfusion in a perioperative period and the prediction of adverse event risks of blood transfusion; and the clinical support result display module integrates the blood use evaluation result and carries out early warning and recommendation according to the blood use evaluation result. Furthermore, the system also comprises a dynamic monitoring module which is used for monitoring the system use log of the user, the feedback of the blood use evaluation result and the decision data of the blood use management process in real time, obtaining evaluation information and sending the abnormal information to the iterative optimization module when the evaluation information has abnormality; the iterative optimization module is triggered periodically or by received abnormal information, the artificial intelligence calculation module is used for training and updating the classification model, if the data expansion requirement exists, the information acquisition module is used for completing data increment, and then newly added data or knowledge is input into the artificial intelligence calculation module for training and updating the classification model and continuous optimization. Therefore, the embodiment of the invention constructs the preoperative-intraoperative-postoperative full-period blood vessel management platform, realizes multi-link coverage and dynamic monitoring of perioperative blood management, and reduces the blood transfusion risk to the maximum extent.
Drawings
Fig. 1 is a schematic structural diagram of a perioperative blood management system according to an embodiment of the present invention.
Fig. 2 is a schematic processing procedure diagram of an information processing module according to an embodiment of the present invention.
Fig. 3 is a schematic processing procedure diagram of an artificial intelligence computing module according to an embodiment of the present invention.
Fig. 4 is a processing diagram of the clinical support result display module according to the embodiment of the present invention.
Fig. 5 is a schematic processing diagram of a dynamic monitoring module according to an embodiment of the present invention.
Fig. 6 is a schematic processing procedure diagram of an iterative optimization module according to an embodiment of the present invention.
FIG. 7 is a flowchart of a perioperative blood management method according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and examples.
In order to realize multi-link coverage and dynamic monitoring of perioperative blood management and reduce blood transfusion risks to the maximum extent, the embodiment of the invention introduces an artificial intelligence engine, evaluates perioperative blood consumption by training an artificial intelligence model, and can optimize the trained artificial intelligence model in the process, so that evaluation is more accurate. Blood use management in perioperative period is carried out according to the blood use evaluation result obtained by the artificial intelligence calculation module, so that the blood transfusion risk can be reduced to the maximum extent.
Specifically, as shown in fig. 1, fig. 1 is a schematic structural diagram of a perioperative blood management system according to an embodiment of the present invention, including: an information acquisition module, an information processing module, an artificial intelligence calculation module and a clinical support result display module, wherein,
the information acquisition module is used for acquiring diagnosis and treatment data of a patient in real time;
the information processing module is used for extracting important blood evaluation features based on diagnosis and treatment data of a patient and converting the important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning;
the artificial intelligence calculation module is used for constructing a classification model based on the characteristics recognizable by the artificial intelligence engine, realizing the assessment of anemia before an operation, the prediction of perioperative blood transfusion and the prediction of risks of adverse events of blood transfusion after the operation, and obtaining a blood use assessment result;
and the clinical support result display module is used for integrating blood use evaluation results and carrying out early warning and recommendation according to the evaluation results.
In the system, further comprising: a dynamic monitoring module and an iterative optimization module, wherein,
the dynamic monitoring module is used for monitoring a system use log of a user, feeding back a blood use evaluation result and decision data of a blood use management process in real time, obtaining evaluation information and sending the abnormal information to the iterative optimization module when the evaluation information is abnormal;
the iterative optimization module is used for triggering periodically or by received abnormal information, the artificial intelligence calculation module is used for training and updating the classification model, if the data expansion requirement exists, the information acquisition module is used for finishing data increment, and then newly added data or knowledge is input into the artificial intelligence calculation module to continuously optimize the system.
Specifically, in the system, the information acquisition module specifically collects structured or unstructured diagnosis and treatment data of a patient;
the information processing module is used for extracting important blood evaluation features on the basis of the diagnosis and treatment data of the patient collected by the information collection module, and converting the extracted important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning;
the artificial intelligence calculation module is trained to obtain a classification model by specifically depending on a machine learning algorithm and a knowledge graph technology, so that assessment of anemia before an operation, prediction of perioperative blood transfusion and prediction of risks of adverse events of postoperative blood transfusion are realized;
the clinical support result display module is used for integrating blood use evaluation results obtained by the artificial intelligent calculation module into a Clinical Decision Support System (CDSS) after being manually confirmed by a transfusion department, displaying a medication recommendation scheme and intraoperative hand anesthesia recommendation operation for an anemia patient, prompting perioperative blood preparation and blood inventory, and early warning for transfusion and high-risk patients with adverse events;
the dynamic monitoring module is used for collecting the artificial confirmation condition of the transfusion department and the adoption condition of clinical personnel in real time, evaluating the application of the CDSS and giving optimization reminding when abnormality occurs;
the iterative optimization module is specifically triggered at regular intervals or when abnormity occurs, newly added data are incorporated into the training process of the classification model, the incorporation characteristics are combined with clinical opinion pairs or the data are added or removed, and the iterative optimization model is continuously iterated, so that the prediction accuracy is improved.
The functions of each module involved in the system are explained in detail below.
Information acquisition module
This module can utilize different systems to collect the data of diagnosing of patient, and specific process includes:
step one, acquiring an identity of a patient, wherein the identity can be a patient hospitalization number or an outpatient number;
collecting the test record of the patient based on the identity of the patient through a Laboratory Information System (LIS);
acquiring operation related data of the patient in the operation through an operation anesthesia information system based on the identity of the patient;
step four, acquiring the imaging examination data of the patient based on the identity of the patient through an image Archiving and Communication system (PACS);
acquiring the past Medical history, the Medical history Record, the diagnosis Record and the medication Record of the patient based on the identity of the patient through an Electronic Medical Record (EMR) and a Hospital Information System (HIS);
and step six, collecting blood consumption information of the patient through a blood distribution system based on the identity of the patient.
Here, the execution order of the second to sixth steps is not limited. And step two to step six use the identity identification association of the patient and integrate different diagnosis and treatment data of the patient acquired in different systems.
Information processing module
As shown in fig. 2, fig. 2 is a schematic view of a processing process of an information processing module according to an embodiment of the present invention, where the specific process includes:
step one, searching a literature using blood vessel management;
step two, determining a data range of blood vessel management;
in the step, according to the search of the literature of the angiology, the risk factors related to the medicine for anemia, blood transfusion and adverse events thereof are determined, and the data range of the angiology is defined;
step three, after unstructured data in the blood management data range are subjected to structuring processing, all structured data are subjected to preprocessing to be converted into a data form which can be identified by an artificial intelligence calculation module;
in this step, the process of performing the structuring process on the unstructured data includes: firstly, the Optical Character Recognition (OCR) technology can be utilized to convert the printed characters in paper documents such as image reports and scanned cases into images composed of black and white dot matrixes, and further to convert the Character Recognition in the images into text data; secondly, by utilizing a Natural Language Processing (NLP) technology, a medical dictionary of diseases, vital signs and symptoms is customized in advance, word segmentation Processing is carried out on unstructured text data, the theme of the segmented text data is extracted, and named entity recognition and relationship judgment are further carried out. And finally, extracting sign values or judging whether diseases exist or not by using the regular expression, and realizing information conversion and structured representation of unstructured data.
In this step, all the structured data are preprocessed to be converted into a data form that can be recognized by the artificial intelligence computing module, and actually, the structured data are coded and mapped. In consideration of the modeling requirement, encoding and mapping are carried out on the classification variables, firstly, synonymy expressions in the classification variables are unified, and then, the multi-classification variables are subjected to dummy variable processing. For example, there are four values for the previous coronary heart disease: presence/absence/yes/no, and merge/yes, which are unified into presence/absence/no, and then are converted into (1, 0) and (0, 1) through the dummy variable processing. And through coding mapping processing, the classification variables are converted into numerical variables which can be identified by an artificial intelligence calculation module.
Artificial intelligence computing module
As shown in fig. 3, fig. 3 is a schematic processing diagram of an artificial intelligence computing module according to an embodiment of the present invention. The artificial intelligence calculation module takes a machine learning algorithm and a knowledge map technology as engines, and trains classification models respectively aiming at different stages of blood vessels by using preprocessed data, and specifically comprises a recommendation model and a prediction model, such as a preoperative anemia medication recommendation model and an intra-operative and postoperative blood transfusion and adverse event prediction model.
For the recommendation model, the characteristics of the previous patient are analyzed, the similarity between the current patient and the previous patient is calculated, the previous patient most similar to the current patient is obtained, candidate items are generated according to the medication or operation of the part of similar patients, and the operation recommendation such as medication of the patient is further perfected by combining a knowledge map. The embodiment of the invention uses a K-nearest neighbor algorithm to judge the position characteristics of the current anemia patient, calculate the distance between the previous anemia patient and the patient, and search K nearest samples in the characteristic space of the patient. According to the label of the anemia-correcting drugs used by k anemia patients, the categories with the highest occurrence frequency are selected as the anemia-correcting drugs of the patients. And manually marking and inducing the entities and the relations in the medicine specification to form an entity and relation rule base. On the basis, a machine learning model is trained, the unlabeled medicine specification is subjected to prediction labeling, and finally a medicine knowledge graph integrating knowledge of medicine indications, contraindications, usage and dosage and the like is constructed. And (3) combining candidate medicines generated based on similar patients, removing contraindicated medicines in the candidate medicines, and providing the finally recommended medicine or medicines with the contents of medicine names, usage amounts and the like.
For the prediction model, during training, the diagnosis and treatment data of a previous patient are divided into a training set and a testing set, a grid search method is used for carrying out exhaustive search on the hyperparameters of the model to establish the model, and the optimal hyperparameter combination and threshold are selected according to the model performance on the testing set. After training, transmitting the characteristics of the current patient into a model to obtain a prediction result, wherein the prediction result comprises the blood transfusion volume, the blood transfusion risk probability value and the adverse event risk probability value, and if the risk probability value is higher than a set threshold value, carrying out early warning prompt. According to the embodiment of the invention, the XGboost algorithm can be used for constructing the prediction model. The XGboost is an integrated algorithm which is based on a series of decision trees and carries out prediction and takes gradient lifting as a framework. The tree-based modeling approach is able to naturally learn the higher order interactions between variables and interpret the non-linear relationships without prior assumptions, and thus may more effectively capture the potentially complex relationships between variables and outcomes. The prediction model provided by the embodiment of the invention comprises perioperative blood transfusion prediction and postoperative blood transfusion adverse event prediction.
Clinical support result display module
Fig. 4 is a processing diagram of the clinical support result display module according to the embodiment of the present invention, as shown in the figure:
and after receiving the blood use evaluation result processed by the artificial intelligence calculation module, pushing the blood use evaluation result to the transfusion department. After confirmation by the transfusion department, the CDSS is integrated, and the early warning result and the recommendation result of the clinical personnel for the branch patients are displayed on the computer side or the mobile side of the clinical personnel. The method specifically comprises the following steps: displaying similar patients and anemia-correcting medication recommendation results aiming at anemia patients before operation; identifying perioperative blood transfusion high-risk patients, reminding clinical personnel to apply for blood preparation and giving a recommended result of hand anesthesia prevention operation; and carrying out high-risk early warning on adverse events of postoperative blood transfusion, and prompting the nursing grade required to be adopted, and the like.
Dynamic monitoring module
Fig. 5 is a schematic processing diagram of a dynamic monitoring module according to an embodiment of the present invention. As shown in the figure, the dynamic monitoring module includes a monitoring data collection unit, an evaluation index calculation unit, and an early warning unit. Wherein the content of the first and second substances,
the monitoring data collection unit is used for collecting and storing the feedback result of the blood transfusion department confirmation step before result display, the collected clinical staff use log data and the patient blood vessel administration decision data in real time;
the evaluation index calculation unit is used for automatically calculating multi-dimensional evaluation indexes including but not limited to accuracy, usability, acceptance rate and the like in real time aiming at different personnel such as clinical personnel, transfusion departments and the like according to the monitoring data collected by the monitoring data collection unit;
and the early warning unit is used for drawing a statistical chart to display the index change in real time and carrying out early warning when any evaluation index is lower than a set threshold value according to the threshold value set by each index.
Iterative optimization module
Fig. 6 is a schematic processing procedure diagram of an iterative optimization module according to an embodiment of the present invention. It can be seen that, depending on the dynamic monitoring module, the iterative optimization module includes a timing iteration unit and a specific condition triggering iteration unit. Wherein the content of the first and second substances,
the timing iteration unit is used for collecting new data generated in a period when the system is used for a certain period, expanding a database and updating and perfecting a knowledge map so as to dynamically optimize the system;
the specific condition triggering iteration unit includes, but is not limited to, evaluating an index triggering early warning such as a great decrease in accuracy, or triggering early warning when a change demand exists for the system in clinic, and once a preset updating condition is reached, adjusting a system module according to a specific condition to complete optimization.
Therefore, multi-link coverage and dynamic monitoring of perioperative blood management can be realized, and blood transfusion risks are reduced to the maximum extent.
Fig. 7 is a flowchart of a perioperative blood management method according to an embodiment of the present invention, which includes the following steps:
701, acquiring diagnosis and treatment data of a patient in real time;
703, constructing a classification model based on the characteristics recognizable by the artificial intelligence engine, and realizing assessment of anemia before an operation, prediction of perioperative blood transfusion and prediction of risks of adverse events of blood transfusion after the operation to obtain a blood use assessment result;
and step 704, integrating blood utilization evaluation results, and performing early warning and recommendation according to the evaluation results.
The method further comprises the following steps: monitoring a system use log of a user in real time, feeding back a blood use evaluation result and decision data of a blood use management process, and obtaining evaluation information;
when the evaluation information has abnormality, acquiring abnormality information;
and triggering periodically or by received abnormal information, inputting newly added data or knowledge according to specific conditions, and performing training updating and continuous optimization on the classification model.
The method and the system provided by the embodiment of the invention can be realized by a computer program, quickly and accurately provide the assistant decision information of blood management for clinical personnel and transfusion department personnel, and construct a novel blood management clinical decision assistant system. The specific implementation process comprises the following steps: after the patient is admitted to a hospital and is determined to be operated, the system is automatically triggered to complete data collection, data processing and artificial intelligence calculation, and a generated result is pushed to a transfusion department for confirmation; if the patient has anemia symptoms, the system gives a reference for drug use for correcting anemia; determining that the operation is started before the operation, periodically predicting blood transfusion risk and blood transfusion amount in the operation, performing early warning when the blood transfusion risk is higher than a certain threshold value, recommending preventive operation of hand anesthesia, displaying an early warning mark when the stock of a transfusion department is lower than the recent blood demand of a patient, and allowing a transfusion department doctor to consider supplementing blood stock related matters; and (4) periodically predicting blood transfusion and adverse event risks thereof after operation, and giving high-risk early warning and nursing grade strengthening prompts. After receiving the prompt, the clinical staff can confirm and modify the prompt, and the background collects the feedback data of the transfusion department and the clinical staff for dynamic monitoring, so that the model updating system is optimized when the triggering condition is met.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A system for perioperative blood management, comprising: an information acquisition module, an information processing module, an artificial intelligence calculation module and a clinical support result display module, wherein,
the information acquisition module is used for acquiring diagnosis and treatment data of a patient in real time;
the information processing module is used for extracting important blood evaluation features based on diagnosis and treatment data of a patient and converting the important blood evaluation features into features which can be recognized by an artificial intelligence engine through data cleaning;
the artificial intelligence calculation module is used for constructing a classification model based on the characteristics recognizable by the artificial intelligence engine, realizing the assessment of anemia before an operation, the prediction of perioperative blood transfusion and the prediction of risks of adverse events of blood transfusion after the operation, and obtaining a blood use assessment result;
and the clinical support result display module is used for integrating the blood use evaluation results and carrying out early warning and recommendation according to the blood use evaluation results.
2. The system of claim 1, wherein the system further comprises: a dynamic monitoring module and an iterative optimization module, wherein,
the dynamic monitoring module is used for monitoring a system use log of a user, feeding back a blood use evaluation result and decision data of a blood use management process in real time, obtaining evaluation information and sending the abnormal information to the iterative optimization module when the evaluation information is abnormal;
the iterative optimization module is used for triggering periodically or by received abnormal information, the artificial intelligence calculation module is used for training and updating the classification model, if the data expansion requirement exists, the information acquisition module is used for finishing data increment, and then newly added data or knowledge is input into the artificial intelligence calculation module to continuously optimize the system.
3. The system of claim 1 or 2, wherein the information acquisition module is further configured to obtain clinical data of the patient in real time, and comprises:
acquiring the identity of a patient; collecting a test record of a patient based on the identity of the patient through a Laboratory Information System (LIS); acquiring operation related data of a patient in an operation based on the identity of the patient through an operation anesthesia information system; acquiring the imaging examination data of the patient based on the identity of the patient through an image archiving and communication system PACS; acquiring the past medical history, the medical course record, the diagnosis and the medication record of a patient based on the identity of the patient through an electronic medical record system EMR and a hospital information system HIS; blood usage information is collected for the patient by the blood dispensing system based on the patient's identification.
4. The system according to claim 1 or 2, wherein the information processing module is further configured to, after structuralizing unstructured data within the blood-managed data range according to the data range determined by the blood-management literature search result, preprocessing all structured data to convert the structuralized data into a data form recognizable by the artificial intelligence calculation module;
the process of carrying out the structuring processing on the unstructured data comprises the following steps: converting printing characters in a paper file into an image formed by a black and white dot matrix by using an Optical Character Recognition (OCR) technology, and then recognizing characters in the image to convert the characters into text data; utilizing a natural language processing NLP technology, customizing a medical dictionary of diseases, vital signs and symptoms in advance, performing word segmentation processing on unstructured text data, extracting a theme of the text data after word segmentation, and further performing named entity recognition and relationship judgment; extracting sign values or judging whether diseases exist or not by using a regular expression, and realizing information conversion and structured representation of unstructured data;
the preprocessing is carried out on all the structured data to convert the structured data into a data form which can be identified by the artificial intelligence calculation module, and the data form is subjected to coding mapping on the structured data.
5. The system of claim 1 or 2, wherein the artificial intelligence computation module further comprises a recommendation model and a prediction model, wherein,
the recommendation model is used for analyzing the characteristics of the previous patient, calculating the similarity between the current patient and the previous patient, acquiring the previous patient most similar to the current patient, generating candidate items according to the medication or operation of the similar patients, and further improving the operation recommendation such as medication of the patient by combining a knowledge graph;
the prediction model is used for splitting diagnosis and treatment data of a previous patient into a training set and a test set during training, performing exhaustive search on hyperparameters of the model by using a grid search method to establish the model, and selecting an optimal hyperparameter combination and a threshold according to model performances on the test set; after training, transmitting the characteristics of the current patient into a model to obtain a prediction result, wherein the prediction result comprises the blood transfusion volume, the blood transfusion risk probability value and the adverse event risk probability value, and if the risk probability value is higher than a set threshold value, carrying out early warning prompt.
6. The system of claim 1 or 2, wherein the clinical support result display module is further configured to integrate the blood use evaluation result obtained by the artificial intelligence calculation module into a CDSS (clinical decision support system) after being confirmed by the transfusionist, display a medication recommendation scheme and an intraoperative hand anesthesia recommendation operation for the anemic patient, prompt perioperative blood preparation and blood inventory, and warn patients with high risk of blood transfusion and adverse events.
7. The system of claim 2, wherein the dynamic monitoring module further comprises: a monitoring data collection unit, an evaluation index calculation unit and an early warning unit, wherein,
the monitoring data collection unit is used for collecting and storing the feedback result of the blood transfusion department confirmation step before result display, the collected clinical staff use log data and the patient blood vessel administration decision data in real time;
the evaluation index calculation unit is used for automatically calculating multi-dimensional evaluation indexes including but not limited to accuracy, usability, acceptance rate and the like in real time aiming at different personnel such as clinical personnel, transfusion departments and the like according to the monitoring data collected by the monitoring data collection unit;
and the early warning unit is used for drawing a statistical chart to display the index change in real time and carrying out early warning when any evaluation index is lower than a set threshold value according to the threshold value set by each index.
8. The system of claim 2, wherein the iterative optimization module further comprises a timed iteration unit and a condition-specific triggered iteration unit, wherein,
the timing iteration unit is used for collecting new data generated in a period when the system is used for a certain period, expanding a database, updating and perfecting a knowledge map and performing dynamic optimization on the system;
and the specific condition triggering iteration unit is used for adjusting the artificial intelligence calculation module according to specific conditions when the evaluation index triggers early warning or clinical change requirements, and if the data expansion requirements exist, the information acquisition module completes data increment and then completes optimization.
9. A method of perioperative blood management, the method comprising:
acquiring diagnosis and treatment data of a patient in real time;
extracting important blood evaluation features based on diagnosis and treatment data of a patient, and cleaning and converting the important blood evaluation features into features recognizable by an artificial intelligence engine;
constructing a classification model based on the recognizable characteristics of an artificial intelligence engine, and realizing assessment of anemia before an operation, prediction of perioperative blood transfusion and prediction of risks of adverse events of blood transfusion after the operation to obtain a blood use assessment result;
and integrating blood evaluation results, and carrying out early warning and recommendation according to the evaluation results.
10. The method of claim 9, wherein the method further comprises:
monitoring a system use log of a user in real time, feeding back a blood use evaluation result and decision data of a blood use management process, and obtaining evaluation information;
when the evaluation information has abnormality, acquiring abnormality information;
and triggering periodically or by received abnormal information, inputting newly added data or knowledge, training and updating the classification model, and continuously optimizing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110986673.9A CN113611401A (en) | 2021-08-26 | 2021-08-26 | Perioperative blood management system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110986673.9A CN113611401A (en) | 2021-08-26 | 2021-08-26 | Perioperative blood management system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113611401A true CN113611401A (en) | 2021-11-05 |
Family
ID=78342093
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110986673.9A Pending CN113611401A (en) | 2021-08-26 | 2021-08-26 | Perioperative blood management system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113611401A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114283947A (en) * | 2021-12-27 | 2022-04-05 | 北京和兴创联健康科技有限公司 | Health management method and system suitable for surgical patients |
| CN116030990A (en) * | 2022-12-26 | 2023-04-28 | 北京和兴创联健康科技有限公司 | Perioperative blood transfusion scheme generation method and system based on cascading model |
| CN116936105A (en) * | 2023-09-18 | 2023-10-24 | 山东朱氏药业集团有限公司 | Intelligent blood sampling regulation and control system based on medical use |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130262357A1 (en) * | 2011-10-28 | 2013-10-03 | Rubendran Amarasingham | Clinical predictive and monitoring system and method |
| CN107799160A (en) * | 2017-10-26 | 2018-03-13 | 医渡云（北京）技术有限公司 | Medication aid decision-making method and device, storage medium, electronic equipment |
| CN111063448A (en) * | 2019-09-06 | 2020-04-24 | 中国人民解放军总医院 | Establishment method, storage system and active early warning system of blood transfusion adverse reaction database |
| CN111724910A (en) * | 2020-05-25 | 2020-09-29 | 北京和兴创联健康科技有限公司 | Detection and evaluation method suitable for blood management of perioperative patients |
-
2021
- 2021-08-26 CN CN202110986673.9A patent/CN113611401A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130262357A1 (en) * | 2011-10-28 | 2013-10-03 | Rubendran Amarasingham | Clinical predictive and monitoring system and method |
| CN107799160A (en) * | 2017-10-26 | 2018-03-13 | 医渡云（北京）技术有限公司 | Medication aid decision-making method and device, storage medium, electronic equipment |
| CN111063448A (en) * | 2019-09-06 | 2020-04-24 | 中国人民解放军总医院 | Establishment method, storage system and active early warning system of blood transfusion adverse reaction database |
| CN111724910A (en) * | 2020-05-25 | 2020-09-29 | 北京和兴创联健康科技有限公司 | Detection and evaluation method suitable for blood management of perioperative patients |
Non-Patent Citations (2)
| Title |
|---|
| 孙波;葛东梅;刘姣;吕翠;程聪;: "智慧输血信息管理系统在围术期血液管理中的应用探讨", 中国输血杂志, no. 05, pages 548 - 550 * |
| 赵一贺;: "术前血栓弹力图预测肝移植术中大量输血的风险", 实用器官移植电子杂志, no. 03, pages 172 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114283947A (en) * | 2021-12-27 | 2022-04-05 | 北京和兴创联健康科技有限公司 | Health management method and system suitable for surgical patients |
| CN116030990A (en) * | 2022-12-26 | 2023-04-28 | 北京和兴创联健康科技有限公司 | Perioperative blood transfusion scheme generation method and system based on cascading model |
| CN116030990B (en) * | 2022-12-26 | 2023-10-27 | 北京和兴创联健康科技有限公司 | Perioperative blood transfusion scheme generation method and system based on cascading model |
| CN116936105A (en) * | 2023-09-18 | 2023-10-24 | 山东朱氏药业集团有限公司 | Intelligent blood sampling regulation and control system based on medical use |
| CN116936105B (en) * | 2023-09-18 | 2023-12-01 | 山东朱氏药业集团有限公司 | Intelligent blood sampling regulation and control system based on medical use |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111986770B (en) | Prescription medication auditing method, device, equipment and storage medium | |
| US20180293354A1 (en) | Clinical content analytics engine | |
| CN113611401A (en) | Perioperative blood management system and method | |
| CN112863630A (en) | Personalized accurate medical question-answering system based on data and knowledge | |
| CN109785927A (en) | Clinical document structuring processing method based on internet integration medical platform | |
| CN109378066A (en) | A kind of control method and control device for realizing disease forecasting based on feature vector | |
| CN111666477A (en) | Data processing method and device, intelligent equipment and medium | |
| CN112614578B (en) | Doctor intelligent recommendation method and device, electronic equipment and storage medium | |
| CN111191415A (en) | Operation classification coding method based on original operation data | |
| CN115579104B (en) | Manual intelligent-based liver cancer whole course digital management method and system | |
| CN115293161A (en) | Reasonable medicine taking system and method based on natural language processing and medicine knowledge graph | |
| CN113782125B (en) | Clinic scoring method and device based on artificial intelligence, electronic equipment and medium | |
| CN111063448A (en) | Establishment method, storage system and active early warning system of blood transfusion adverse reaction database | |
| CN114334175B (en) | Hospital epidemic situation monitoring method and device, computer equipment and storage medium | |
| CN110164519B (en) | Classification method for processing electronic medical record mixed data based on crowd-sourcing network | |
| Jiang et al. | Stroke risk prediction using artificial intelligence techniques through electronic health records | |
| CN116775897A (en) | Knowledge graph construction and query method and device, electronic equipment and storage medium | |
| CN115248842A (en) | ICD intelligent coding system based on knowledge graph and retrieval engine | |
| Manoharan | Leveraging machine learning and NLP for enhanced cohorting and RxNorm mapping in Electronic Health Records (EHRs) | |
| CN116884612A (en) | Intelligent analysis method, device, equipment and storage medium for disease risk level | |
| WO2023217737A1 (en) | Health data enrichment for improved medical diagnostics | |
| CN115036034B (en) | Similar patient identification method and system based on patient characterization map | |
| CN114664421A (en) | Doctor-patient matching method and device, electronic equipment, medium and product | |
| CN114613477A (en) | Psychological monitoring method, device, equipment and storage medium based on deep learning | |
| CN112530598A (en) | Health risk self-measurement table recommendation method and system based on health data |
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 |