Preoperative anesthesia assessment intelligent management system and preoperative anesthesia assessment intelligent analysis method

Abstract

The invention provides an intelligent management system for preoperative anesthesia assessment, and the system comprises an online APP patient end which is used for a patient to carry out preoperative self-assessment and information interaction on mobile terminal equipment; the online APP doctor end is used for medical staff to check the preoperative assessment information of the patient on mobile terminal equipment and give an anesthesia risk assessment result; the offline self-service machine end is used for providing self-service evaluation service and information interaction for offline patients; and the offline PC doctor terminal is used for medical staff to check the preoperative assessment information of the patient in a hospital or a workstation and give an anesthesia risk assessment result. Furthermore, the invention further discloses an intelligent analysis method for preoperative anesthesia evaluation. Through the operation platform integrating multiple terminals and multiple scenes, full-process management from patient self-evaluation, doctor evaluation, agreement signing to report uploading is achieved, a standardized, intelligent and paperless evaluation closed loop is formed, and the efficiency and accuracy of preoperative anesthesia evaluation are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of medical information technology, in particular to a preoperative anesthesia evaluation intelligent management system and a preoperative anesthesia evaluation intelligent analysis method. BACKGROUND

[0002] In modern medical treatment, anesthesia is an important link in the surgical process and an important guarantee for the success of surgery. However, the anesthesia process itself also has certain risks, especially for children, the elderly and patients with comorbidities, the anesthesia risk is higher. Therefore, preoperative anesthesia evaluation is particularly important. The traditional anesthesia evaluation method mainly relies on face-to-face communication between medical staff and patients, manually inquiring the patient's medical history, allergy history, physical examination and other information, relying on the experience of medical staff to evaluate the anesthesia risk of patients, and formulating relevant anesthesia programs. However, this method has many problems, mainly including: (1) low evaluation efficiency, long patient waiting time, and poor medical experience; (2) strong subjectivity of anesthesiologists, making it difficult to ensure the objectivity and accuracy of the evaluation; (3) the data collected in the evaluation process are scattered in different paper records or systems, and the data is difficult to manage uniformly, resulting in incoherent data and serious information island problem; (4) lack of standardized process, leading to non-standard evaluation process, especially for special groups (such as children, the elderly, chronic disease patients, etc.) The evaluation is more likely to have problems; (5) low degree of informatization, paper questionnaire, signature confirmation and other methods are still used, the evaluation process is cumbersome and prone to errors, and it is difficult to realize information management. SUMMARY

[0003] To solve the problems of traditional anesthesia evaluation methods, such as the need for face-to-face communication with patients, low evaluation efficiency, strong subjectivity, non-standard process, incoherent data, and insufficient informatization, the present application provides a preoperative anesthesia evaluation intelligent management system, which aims to provide comprehensive and intelligent preoperative anesthesia evaluation management tools for patients, anesthesiologists and hospital managers through multiple terminals and multiple scene integration. The system integrates patient self-evaluation, doctor review, intelligent analysis, data integration and other functions, and realizes standardized, intelligent and paperless management of anesthesia evaluation through multi-platform operation, improves the efficiency and accuracy of preoperative anesthesia evaluation, and provides reliable decision support for clinical anesthesia. Further, the present application also discloses a preoperative anesthesia evaluation intelligent analysis method, which can provide anesthesia risk evaluation guidance score value for doctors to refer to anesthesia risk evaluation results, making the anesthesia risk evaluation results given by medical staff more objective and scientific.

[0004] The application discloses a preoperative anesthesia evaluation intelligent management system, which comprises an online Internet end and an offline hospital end; the online Internet end comprises an APP patient end, which is used for preoperative self-evaluation and information interaction of patients on a mobile terminal device; and an APP doctor end, which is used for medical staff to check preoperative evaluation information of patients on the mobile terminal device and give anesthesia risk evaluation results; the offline hospital end comprises a self-service machine end, which is used for providing self-service evaluation service and information interaction for offline patients; and a PC doctor end, which is used for medical staff to check preoperative evaluation information of patients in a hospital or a workstation and give anesthesia risk evaluation results; the content of the preoperative evaluation information comprises basic information of patients, medical history information, photographed image information and recorded video information; the anesthesia risk evaluation results comprise passing and needing offline consultation; the APP patient end and the APP doctor end are connected in communication through an Internet server, the self-service machine end and the PC doctor end are connected in communication through a hospital server, and the Internet server and the hospital server are connected in communication.

[0005] Further, the APP patient end comprises the following function modules: a medical history collection module for filling in preoperative evaluation information and uploading the filled preoperative evaluation information to an Internet server; an evaluation result acquisition module for extracting the anesthesia risk evaluation result made by a doctor from an anesthesia evaluation report, prompting the patient to confirm the anesthesia risk evaluation result, and pushing an anesthesia informed consent form when the anesthesia risk evaluation result is "pass", and uploading the signature data after the patient performs electronic signature confirmation to an Internet CA signature server; a report preview module for obtaining the anesthesia evaluation report generated by the APP doctor end from the Internet server and showing the anesthesia evaluation report to the patient; the APP doctor end comprises the following function modules: a management audit module for obtaining the preoperative evaluation information of the patient from the Internet server, generating an anesthesia evaluation report based on the anesthesia risk evaluation result given by the doctor and uploading it to the Internet server, and pushing an anesthesia informed consent form when the anesthesia risk evaluation result is "pass", and uploading the signature data after the medical staff performs electronic signature confirmation to the Internet CA signature server; a history record module for obtaining the anesthesia evaluation report from the Internet server and archiving and storing the anesthesia evaluation report; the self-service machine end comprises all the function modules of the APP patient end, the only difference being that the medical history collection module uploads the filled preoperative evaluation information to an intranet server; the evaluation result acquisition module uploads the signature data after the patient performs electronic signature confirmation to an intranet CA signature server; the report preview module obtains the anesthesia evaluation report generated by the APP doctor end from the intranet server; the self-service machine end further comprises a patient identity recognition module for recognizing the user identity and obtaining the patient related information; the PC doctor end comprises all the function modules of the APP doctor end, the only difference being that the management audit module obtains the preoperative evaluation information of the patient from the intranet server, uploads the anesthesia evaluation report to the intranet server, and uploads the signature data after the medical staff performs electronic signature confirmation to the intranet CA signature server; the history record module is used to obtain the anesthesia evaluation report from the intranet server.

[0006] Further, the APP patient end and the self-service machine end further comprise an intelligent guidance module; the intelligent guidance module is used to provide animation guidance and voice prompts during the preoperative evaluation information filling process to help the patient complete the preoperative self-evaluation.

[0007] Further, the intelligent guidance module can also convert the recorded video information of the patient's corresponding actions according to the animation guidance and voice prompts in the preoperative evaluation information into quantitative data information through an artificial intelligence algorithm.

[0008] Further, the APP doctor end and the PC doctor end further comprise an intelligent analysis module; the intelligent analysis module is used to intelligently analyze the preoperative evaluation information and give a preoperative anesthesia risk evaluation guidance score value.

[0009] The intelligent analysis module specifically comprises: a data preprocessing unit, configured to automatically extract text, picture and video data in preoperative assessment information, and pre-process the extracted data of different modalities respectively; a feature extraction unit, configured to perform feature extraction on the pre-processed text data and image data respectively to obtain features of different modalities of data; a feature fusion unit, configured to splice, reduce dimension and pay attention to the features to obtain fused features; and a prediction output unit, configured to convert the fused features into a score value of anesthesia risk assessment guidance provided for reference by a doctor.

[0010] Further, the management and auditing module has at least one of the following functions: the management and auditing module is further configured to collect and arrange the submitted preoperative assessment information, and sort the preoperative assessment information according to submission time or patient priority; the management and auditing module has a priority sorting function, and is configured to sort the patient standardized forms according to operation dates or emergency degrees of diseases; and the management and auditing module has a doctor-patient online communication function, and supports online communication between doctors and patients.

[0011] Further, the PC doctor terminal further comprises a statistical analysis module; the statistical analysis module is configured to perform multi-dimensional statistics and analysis on the assessment data, customize individualized reports, visually display data, and review historical data.

[0012] Further, the PC doctor terminal further comprises a system management module; the system management module is configured to perform system parameter configuration, user permission management, system backup and recovery, log management and data security management.

[0013] Further, the PC doctor terminal and the self-service machine terminal are directly connected in communication, and are configured to import relevant information of a patient registered for a preoperative anesthesia assessment outpatient service into the anesthesia assessment intelligent management system; and the PC doctor terminal is further connected in communication with at least one of a LIS system and a PACS system, and is configured to obtain examination and inspection data of the patient.

[0014] The second aspect of the present application discloses an intelligent preoperative anesthesia assessment analysis method, comprising:

[0015] Obtaining preoperative assessment information filled by an online or in-hospital patient; the content of the preoperative assessment information comprises patient basic information, medical history information, photographed image information and recorded video information;

[0016] Extracting text, picture and video data in the preoperative assessment information, and pre-processing the extracted data of different modalities respectively;

[0017] Performing feature extraction on the pre-processed text data and image data respectively to obtain features of different modalities of data;

[0018] The features are spliced, dimensionally reduced, and attention outputted to obtain fused features;

[0019] The fused features are converted into an anesthesia risk assessment guidance score value provided for reference by a doctor.

[0020] Based on the preoperative anesthesia assessment intelligent analysis method, medical staff checks and gives an anesthesia risk assessment result based on the preoperative assessment information and the anesthesia risk assessment guidance score value at a hospital or workstation; the anesthesia risk assessment result includes pass and needs offline consultation.

[0021] The present application has the following beneficial effects:

[0022] (1) The present application realizes the interconnection and intercommunication of online and offline through the all-round integration of the online APP patient end, the APP doctor end, the offline self-service machine end, and the PC doctor end, and builds a comprehensive intelligent assessment management platform. Patients can conveniently complete preoperative self-assessment, doctors can check and give assessment results online in real time, realize the whole process management from patient self-assessment, doctor assessment, signature of consent form to report uploading, form a standardized, intelligent and paperless assessment closed loop, and greatly improve the assessment efficiency and convenience.

[0023] (2) The intelligent analysis module built in the preoperative anesthesia assessment intelligent management system of the present application can quantitatively process the assessment information of the patient, reduce the error of the subjective judgment of the doctor, and improve the objectivity and accuracy of the assessment result. The module also provides an assessment guidance score, so that the doctor can better master the anesthesia risk situation of the patient.

[0024] (3) The preoperative anesthesia assessment intelligent management system provided by the present application adopts a standardized form to uniformly collect and manage patient information, ensures the standardization of the assessment process; the system also realizes paperless management through an electronic signature function, guarantees information security and compliance, and reduces the tediousness and errors caused by paper operation.

[0025] (4) The preoperative anesthesia assessment intelligent management system provided by the present application also develops an intelligent guidance module for animation and voice prompt to help patients complete the assessment, especially for the elderly and special groups, which not only improves the operation experience of the patient, but also enhances the smoothness and friendliness of the assessment.

[0026] (5) The preoperative anesthesia assessment intelligent management system provided by the present application sets a statistical analysis module on the PC doctor end to analyze the assessment data from multiple angles, generate personalized reports, and provide more comprehensive clinical decision support for doctors, further improving the fine level of assessment management.

[0027] (6) The preoperative anesthesia evaluation intelligent management system provided by the application is also in communication connection with an Internet server, an in-hospital server, an HIS system, an electronic signature system and the like, ensuring efficient transmission and integration of information at various terminals, avoiding the problem of data islands, effectively optimizing the preoperative evaluation process, and further improving the standardization, intelligentization and paperless level of evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 It is a schematic diagram of the architecture of the preoperative anesthesia evaluation intelligent management system.

[0029] Figure 2 It is a flowchart of the implementation of the intelligent analysis module.

[0030] Figure 3 It is a schematic diagram of the multi-modal artificial intelligence algorithm. DETAILED DESCRIPTION

[0031] The application breaks the traditional offline face-to-face evaluation process mode, integrates advanced information technology and clinical anesthesia evaluation requirements, and provides a multi-terminal and multi-scene preoperative anesthesia evaluation intelligent management system, including an APP patient end, an APP doctor end, a self-service machine end and a PC doctor end. The patient can perform self-evaluation through the APP patient end or the self-service machine end in the hospital, fill in the preoperative anesthesia evaluation form, and input personal basic information, medical history information and the like. The system collects these information in the form of a standardized electronic form and uploads them to the hospital central data system (HIS system) in real time. The medical staff can check the preoperative self-evaluation data of the patient through the APP doctor end and the PC doctor end, refer to the anesthesia risk evaluation score given by the system, and give the anesthesia risk evaluation result. The application also integrates electronic signature and intelligent analysis functions, significantly improves the efficiency of the anesthesia evaluation process, the reliability of the data and the accuracy of the decision, provides a more efficient, safe and standardized solution for the clinical anesthesia work and management work of the hospital, and promotes the digitalization and standardization development of medical management.

[0032] The technical solutions of the application will be further explained and described in combination with specific embodiments and the drawings.

[0033] In combination with Figure 1 the drawings, the embodiment of the application provides a preoperative anesthesia evaluation intelligent management system, which includes an online Internet end and an offline in-hospital end. The online Internet end mainly includes an APP patient end and an APP doctor end, and the offline in-hospital end mainly includes a self-service machine end and a PC doctor end.

[0034] The APP patient end is a main entrance for a patient to perform preoperative self-evaluation and information interaction, and is convenient for the patient to fill in anesthesia evaluation information by a mobile terminal device such as a mobile phone or a tablet computer. Specifically, the APP patient end mainly includes a medical history collection module, an intelligent guidance module, an evaluation result acquisition module and a report preview module.

[0035] The medical history collection module is one of core function modules of the APP patient end, and the patient fills in anesthesia evaluation information in detail through the module. The module adopts a standardized form, and anesthesia evaluation information is presented through the standardized form. The content of the standardized form mainly includes: patient basic information such as name, gender, age, height, weight, BMI, smoking history, drinking history and the like; medical history information such as tooth evaluation, whether there is snoring, whether there is hypertension, whether there is heart disease, cold history in the past two weeks, bronchitis / pneumonia history in the past month, asthma history, thyroid disease history, diabetes history, whether there is liver disease, whether there is kidney disease, allergy history, previous anesthesia history, antihypertensive drugs, blood glucose control drugs, anticoagulant / antiplatelet drugs, glucocorticoids, tumor treatment, drug abuse dependence history, Marfan syndrome and the like; image information such as medical reports, test results and the like; video information mainly generated after corresponding actions (such as: opening mouth, tilting head, turning neck left and right, making "ah" sound) are made according to built-in animation guidance and voice prompts in the system. After the above information collection is completed, the standardized form is uploaded and stored to an Internet server as a basis for preoperative anesthesia evaluation. In order to improve the filling experience, the module also supports data saving function, and the patient can fill in stage by stage and return to supplement information when necessary.

[0036] The intelligent guidance module has animation guidance and voice prompt functions for the patient, and helps the patient to complete the standard form filling process. Specifically, the module can intelligently guide the process according to the options clicked / characters input by the patient, help the patient to provide more complete and accurate patient basic information, medical history information and the like; the module also has animation prompt (including animation guidance and voice prompt) and information integration functions, which are used to guide the patient to make airway evaluation related actions (such as: opening mouth, tilting head, turning neck left and right, making "ah" sound and the like) to complete the process of image information shooting and video information recording; further, the module can also convert the image information such as mouth opening and head tilting into quantified data information such as patient mouth opening degree and head and neck activity degree through artificial intelligence algorithms such as face feature recognition and point cloud analysis, and automatically fill the calculation results into the standardized form, so that the standardized form information is easy to read and convenient for doctors to make comprehensive and accurate evaluation. Thus, the patient form filling operation is more simple and humanized, and it is convenient for doctors to make more comprehensive evaluation.

[0037] The evaluation result acquisition module has the functions of extracting the anesthesia risk evaluation result made by the doctor and prompting the patient to confirm the evaluation result. This module can automatically extract the anesthesia risk evaluation result in the anesthesia evaluation report. If the anesthesia risk evaluation result is "pass" state, the patient can perform electronic signature on the anesthesia informed consent form pushed by the module. After the signature is completed, the module uploads the signed data to the third-party system CA electronic signature server storage by calling the CA signature interface.

[0038] The report preview module can obtain the anesthesia evaluation report from the Internet server and show the anesthesia evaluation report to the patient. The content of the anesthesia evaluation report mainly includes patient basic information, medical history information, image information, video information, anesthesia risk evaluation result and other information. The display mode of the anesthesia evaluation report is clear combination of text and images. The module also has the functions of downloading or exporting the anesthesia evaluation report as a PDF file for future reference or sharing with other medical institutions, and multiple viewing of historical anesthesia evaluation reports, which is convenient for patients and doctors to review the evaluation results at any time.

[0039] The APP patient end integrates the functions of medical history collection, report preview, evaluation result acquisition and intelligent guidance through the above-mentioned modules, greatly facilitates the patient to perform preoperative anesthesia self-evaluation, and provides the doctor with perfect anesthesia evaluation data. Moreover, the module improves the evaluation efficiency through the combination of informatization and intelligentization, reduces the communication burden of the patient and the doctor, and realizes the seamless connection of the evaluation process.

[0040] The APP doctor end is the main entrance for the doctor to review the patient's self-evaluation anesthesia evaluation record. The doctor can log in the system through mobile terminal devices such as mobile phones and tablets at any time and any place, access the data stored in the Internet server, view and review the standardized form submitted by the patient, and give the anesthesia risk evaluation result. Specifically, the APP doctor end mainly includes three parts of intelligent analysis module, management audit module and historical record module.

[0041] The intelligent analysis module has a function of anesthesia risk assessment. Specifically, the module realizes the intelligent analysis function based on a multi-modal artificial intelligence algorithm, and the multi-modal artificial intelligence algorithm includes a data preprocessing unit, a feature extraction unit, a feature fusion unit, and a prediction output unit. The data preprocessing unit automatically extracts text information (such as patient basic information and medical history information) in the standardized form, and performs discretization processing on the extracted text information to obtain numerical data, and then performs standardization processing on the numerical data; the picture information (photographed image information) in the standardized form is scaled, and then the scaled image data is standardized; the video information (recorded video information) in the standardized form is extracted to obtain key image frames. The feature extraction unit extracts the effective feature representation of the different modal data output by the data preprocessing unit, converts the preprocessed data into a low-dimensional feature vector suitable for subsequent processing, and obtains data features of different modalities. The feature fusion unit integrates the data features of different modalities into a unified representation through input splicing, vector dimension reduction, and multi-head attention layers, obtains fusion features, and the fusion features contain information from different modalities, which can effectively improve the performance and robustness of the model in comprehensive data analysis. The prediction output unit converts the fusion features into an anesthesia risk assessment guidance score value (for example: a score of more than 80 points, a risk level of 1, defined as a normal airway; a score between 60-80 points, a risk level of 2, defined as a potential difficult airway; a score less than 60 points, a risk level of 3, defined as a difficult airway), to help doctors judge the potential risk of preoperative anesthesia and obtain more accurate anesthesia risk assessment results. For high-risk patients, the system will issue an alarm and prompt the doctor to review the evaluation content in more detail. The module also has a visualization function, which clearly displays the anesthesia evaluation abnormal indicators of each patient in the form of charts and risk levels, helping doctors make quick decisions.

[0042] The management review module has functions of displaying / reviewing the standardized form, anesthesia risk assessment, adding remarks, automatically pulling and pushing data, uploading anesthesia evaluation reports, priority sorting, and online communication between doctors and patients.

[0043] Specifically, the module can collect and organize all the submitted patient standardized forms, and sort them in order according to the submission time or patient priority, etc., to facilitate doctors to quickly understand the tasks to be evaluated. Doctors can view the basic information (such as name, age, operation time, etc.) of the patient in a simple version in the to-be-reviewed list, and enter the detailed evaluation data page by clicking on an item. In the evaluation data page, doctors can view the complete patient basic information, medical history information, photographed image information, and recorded video information.

[0044] The module also has a function of anesthesia evaluation. A doctor can manually complete anesthesia risk evaluation, and can add notes during the review process, or edit and supplement the anesthesia risk evaluation results that have been completed but not formally submitted, to ensure that the final anesthesia evaluation report is accurate and reliable. For anesthesia evaluation reports with an evaluation result of "pass", the doctor can perform electronic signature on the anesthesia informed consent form pushed by the module. After the signature is completed, the module uploads the signed data to the third-party system CA electronic signature server storage by calling the CA signature interface. If the anesthesia risk evaluation result is "not passed", a notification of "need to supplement relevant examination and re-visit, or need offline consultation" is given. The final anesthesia evaluation report is uploaded to the Internet server and stored in JSON format.

[0045] The module also has a function of automatic data pulling and pushing. The module can pull the patient information of the patient registered for anesthesia evaluation outpatient service from the HIS system (hospital central data system) through the ESB interface provided by the HIS system, and import the patient information into the anesthesia evaluation intelligent management system, to reduce the workload of repeated entry. After the patient completes the submission of the standardized form, the module can automatically pull the information to the module and remind the doctor to handle it as soon as possible. After the doctor completes the review, the system will automatically push the evaluation result to the patient end.

[0046] The module also has a function of priority sorting. The doctor can prioritize the evaluation tasks of some patients according to the operation date or the emergency degree of the disease, to ensure an efficient evaluation process. The module also has a function of online communication between doctors and patients, to support online communication between doctors and patients, and to meet the requirements of doctors to ask patients for supplementary materials or preoperative suggestions during the review process.

[0047] The history record module has a function of checking completed anesthesia evaluation reports. Specifically, the history record module can obtain anesthesia evaluation reports that have been reviewed by a doctor in real time through an Internet server, and automatically classify and store the anesthesia evaluation reports according to dates. The doctor can retrieve the historical evaluation records of the patients who have been treated through the module, and view the basic information, medical history information, image information, video information, anesthesia risk evaluation results, and other information of the patients. The module also has a function of screening and searching. The doctor can quickly retrieve the required historical evaluation records according to the patient's name, evaluation time, operation type, and other conditions. In addition, the history record module also supports multiple reviews and edits. The doctor and the hospital can review or supplement and modify the evaluation results that have been reviewed.

[0048] The APP doctor end provides a high-efficiency and convenient preoperative anesthesia evaluation management platform for doctors by integrating the intelligent analysis module, the management audit module, and the historical record module. Through the cooperative work of the three functional modules, the efficiency and accuracy of the doctors in handling anesthesia evaluation tasks are significantly improved, and the overall process of preoperative anesthesia evaluation is greatly optimized.

[0049] The self-service machine end can provide self-service evaluation services and information interaction for offline patients. The self-service machine attached to the self-service machine end is usually deployed in the public area of the hospital, such as the patient self-service area, to help patients complete preoperative evaluation without the intervention of medical personnel. Specifically, the self-service machine end mainly includes a patient identity recognition module, a medical history collection module, an intelligent guidance module, and a report preview module.

[0050] The patient identity recognition module is the entry module of the self-service machine end and has an identity recognition function. Specifically, the module supports multiple identity recognition methods, including card swiping (e.g., medical insurance card, ID card) and code scanning (e.g., medical insurance QR code, medical code). Through these identity recognition information, the ESB interface provided by the HIS system can be automatically called to obtain patient-related information (e.g., patient personal information such as patient HIS system unique ID, name, gender, age), reducing the trouble of repeated information filling and further speeding up the evaluation process.

[0051] The medical history collection module, the intelligent guidance module, the evaluation result acquisition module, and the report preview module have basically the same function as the online internet end medical history collection module, which will not be described here. It should be noted that, unlike the APP patient end, the self-service machine end is connected to the in-hospital server, so each functional module thereof is in communication connection with the in-hospital server to upload standardized forms or download anesthesia evaluation reports and other data.

[0052] The self-service machine end provides a full-process self-evaluation solution for patients by integrating the patient identity recognition module, the medical history collection module, the intelligent guidance module, and the report preview module, which basically requires no manual intervention, not only improves the efficiency and accuracy of the evaluation, but also greatly improves the patient's use experience, realizing the intelligentization and standardization of preoperative anesthesia evaluation.

[0053] The PC doctor end is an important part of the preoperative anesthesia evaluation intelligent management system, mainly for anesthesiologists to use, supports doctors to handle anesthesia evaluation related work in the hospital or workstation, and provides anesthesia risk assessment, historical record viewing, data statistical analysis, and system management functions. Through a more refined management interface and powerful data processing capability, the overall control of the entire evaluation process is realized. The PC doctor end mainly includes an intelligent analysis module, a management audit module, a historical record module, a statistical analysis module, and a system management module.

[0054] The intelligent analysis module, the management audit module, and the historical record module have basically the same functions as the online Internet-side intelligent analysis module, which will not be described here. It should be noted that, unlike the APP doctor end, the PC doctor end is connected to the hospital server. The PC doctor end communicates with the HIS system, the hand anesthesia system, etc. through the hospital server to realize data sharing of the anesthesia evaluation report. At the same time, the PC doctor end can also directly communicate with the HIS, LIS (laboratory information system), PACS (medical imaging system) and other systems to obtain the existing patient personal information, examination data, etc. in these systems.

[0055] The statistical analysis module has functions such as data analysis, report generation, and data review. Specifically, this module can perform multi-dimensional statistics and analysis on evaluation data, monitor the evaluation progress of each department in real time, find bottlenecks in the evaluation process, and adjust resource allocation in a timely manner. This module supports personalized report customization, users can select different analysis dimensions according to their needs, such as patient age, medical history type, surgery type, etc., to generate reports that meet specific needs, and also supports visual display of data, presenting complex statistical data through charts, curves, etc. to facilitate managers to quickly understand and analyze. This module also has a historical data review function, which can help hospitals track postoperative outcomes and analyze the relationship between different types of anesthesia evaluation programs and surgical outcomes, thereby providing data support for future evaluation and treatment.

[0056] The system management module has functions such as system parameter configuration, user permission management, system backup and recovery, log management, and data security management. Specifically, this module is used to maintain the normal operation of the entire preoperative anesthesia evaluation intelligent management system and the configuration of various system parameters. This module allows administrators to allocate system permissions according to different job responsibilities, ensuring that only authorized personnel can access specific functions and data. Administrators can create, edit and delete user accounts, and record user operations throughout the process to ensure the traceability and security of the system. This module supports regular data backup and recovery to ensure that data can be quickly restored in the event of system failure, reducing interruptions to the evaluation process. The log management function allows administrators to view system operation logs and error reports to identify potential problems and fix them in a timely manner. This module also provides data encryption and security protection functions to ensure the security of patient information during transmission and storage.

[0057] The PC doctor end integrates the above modules to realize centralized management of patient information, intelligent anesthesia risk assessment, comprehensive system management, and powerful data statistical analysis, providing doctors with a comprehensive, accurate, and efficient preoperative anesthesia evaluation intelligent management platform.

[0058] It can be understood that the online Internet server and the offline hospital server are in communication connection, data is transmitted through the way of HTTP interface calling, data synchronization is realized, and the sharing of standardized forms and anesthesia evaluation reports and other information is achieved. The hospital server is also in communication connection with the hospital HIS system, and the business data interconnection is realized through the hospital ESB interface, and then the data sharing of anesthesia evaluation reports and other related information in various systems in the hospital (for example, the HIS system, the hand anesthesia system, etc.) is realized.

[0059] Based on the preoperative anesthesia evaluation intelligent management system provided in the embodiments of the application, the online anesthesia evaluation process mainly includes the following steps:

[0060] Firstly, after the patient registers through the APP patient terminal in the Internet hospital, the patient enters the patient treatment interface, fills in the standardized form through the anesthesia evaluation button, and completes the filling of the standardized form. The content filled in includes patient basic information, medical history information, and image information, video information, and examination report uploading. The whole process can be completed under the guidance of the intelligent guidance module, and the final standardized form is formed.

[0061] After that, the data automatic pulling and pushing function of the APP doctor terminal management audit module is started, the patient registered for anesthesia evaluation Internet outpatient service is grabbed through the ESB interface provided by the HIS system, and is imported into the anesthesia evaluation intelligent management system, reminding the doctor to handle it as soon as possible. In addition, after multiple patients register, the priority sorting function of the management audit module is started, and the doctor can prioritize the evaluation tasks of some patients according to the registration order, operation date or disease emergency degree, to ensure efficient evaluation process.

[0062] Then, the intelligent analysis module of the APP doctor terminal is started, which gives a anesthesia risk evaluation guidance score value based on the pulled standardized form through multi-modal artificial intelligence algorithm for intelligent analysis of patient basic information, medical history information, photographed image information, and recorded video information. The analysis steps of the data preprocessing unit, feature extraction unit, feature fusion unit and prediction output unit included in the multi-modal artificial intelligence algorithm are as follows:

[0063] The data preprocessing unit automatically extracts the data information in the standardized standardized form, and respectively pre-processes the extracted text, picture and video data, as follows:

[0064] For the text information (such as patient basic information, medical history information, etc.) in the form: first, the text information is discretized to obtain numerical data; then, the numerical data is standardized according to the mean and standard deviation, so as to conform to the standard normal distribution, and the standardization formula is as follows:

[0065]

[0066] where x is the original numerical data, μ is the mean of the feature, σ is the standard deviation of the feature, x norm is the normalized numerical data.

[0067] For the picture information (such as the photographed image information) in the form: first, the picture information is scaled, and the image data is uniformly sized as HxWxC, H, W and C respectively represent the length, width and channel number of the image, for example, 640*640*3; then the image data is standardized, and the pixel value is scaled to the range of [0, 1], that is, normalized.

[0068] For the video information (such as the recorded video information) in the form: first, the video is converted into a sequence of image frames; then the mean square error (MSE) between the frames is calculated, and when the mean square error is greater than a preset value, the current image frame is extracted as a key image frame, wherein the calculation formula of the mean square error is as follows:

[0069]

[0070] where I1(i) and I2(i) respectively represent the gray value of the i-th pixel of the adjacent two frames of images, and N is the total number of pixels of the image.

[0071] The feature extraction unit respectively extracts features from the preprocessed text data and image data, as follows:

[0072] For the preprocessed text data: the multi-layer perception (MLP) is used to extract the text features. Specifically, the numerical features obtained after standardization are converted into embedding vectors through the multi-layer perception (MLP), and the formula is as follows:

[0073] h = ReLU (Wx norm +b)

[0074] where x norm is the normalized numerical data, W and b are the weight matrix and bias term of linear transformation, ReLU is the activation function, and h is the generated embedding vector.

[0075] Then, embedding is generated for all numerical features, and then they are concatenated to form a unified embedding representation, and the formula is as follows:

[0076] h final = concat (h height , h weight , h age ...)

[0077] where h height , h weight , h ageThe embedded vectors representing height, weight and age, respectively, and the omitted part of the ellipsis is the embedded vector of other text information, h final is the final comprehensive embedded representation. This unified embedded representation h final will be the input of the subsequent functional units, helping the algorithm to understand the role of the text information filled in by the patient in the risk assessment.

[0078] For the preprocessed image data: the ResNet50 model is used to extract image features from the preprocessed image. The ResNet50 model is divided into four parts. The first part consists of a convolutional layer with a convolution kernel size of 7*7.

[0079] The second part consists of four residual block groups containing 3, 4, 6, and 3 residual modules in sequence. Each residual block contains 3 convolutional layers, and the implementation formula of the residual module is as follows:

[0080] H(x) = ReLU(F(x) + x)

[0081] Where F(x) is the output of the convolutional layer combination, which is directly added to the input x.

[0082] The third part consists of an average pooling layer, which pools the features of each channel to generate a feature vector. The calculation formula of the average pooling layer is as follows:

[0083]

[0084] X ijk represents the value of the i-th channel at position (j, k), f i is the average value of each channel.

[0085] The fourth part consists of a fully connected layer, which performs the final classification based on the output of the global pooling.

[0086] The feature fusion unit first fuses the data features, which specifically includes:

[0087] First, the text and image features are concatenated in the last dimension to obtain a joint feature vector. The concatenated joint feature vector f concat = concat(f text , f image ), f text is the text feature vector, and f image is the image feature vector.

[0088] Second, to reduce the dimension of the concatenated feature vector, a fully connected layer can be used to further process the joint feature vector, making it more representative and reducing redundant information. The formula for dimension reduction using a fully connected layer is as follows:

[0089] freduced = σ(W reduced concat + b reduce )

[0090] where W reduced is the weight matrix of the dimension reduction linear layer, b reduce is the bias term, σ is the activation function, and f reduced is the reduced fusion feature vector.

[0091] Thirdly, the attention weight is calculated based on the reduced fusion feature vector f reduced , which measures the contribution of the feature at a certain position in f reduced to the feature at the target position. The attention weight is calculated as follows: for the input feature vector f reduced , first calculate the query matrix Q = f reduced W Q , the key matrix K = f reduced W K , and the value matrix V = f reduced W V , where W Q , W K , and W V are learnable weight matrices used to generate query, key, and value vectors; then calculate the similarity between query and key through scaled dot-product attention to obtain the attention weight.

[0092] Fourthly, the value matrix is weighted and summed using the attention weight to obtain the single-head attention output, calculated as follows:

[0093]

[0094] where head represents the single-head attention output, Q is the query matrix representing the current input to be focused on, K is the key matrix representing the feature representation at each position, V is the value matrix representing the original feature value of the input, d k represents the dimension of the key vector, used for scaling to prevent gradient instability caused by large dot product values, and softmax represents the activation function used to convert similarity scores into probability distributions.

[0095] Fifthly, multiple heads are used to calculate attention in parallel, i.e., the results of each single-head attention output are concatenated and then linearly transformed to generate the final output, calculated as follows:

[0096] MultiHead(Q,K,V) = concat(head1,...,head h )W O

[0097] ​wherein MultiHead(Q, K, V) is the multi-head attention output, containing information of all attention heads, each head1 represents a single-head attention head output, calculated using different weight matrices, W O is the linear transformation weight matrix of the multi-head attention.

[0098] The prediction output unit first performs global average pooling on the multi-head attention output obtained by the feature fusion unit to generate a fixed-length feature vector f feature , and the calculation formula is as follows:

[0099]

[0100] wherein N is the sequence length of the multi-head attention output.

[0101] Then the feature vector f feature is sent to a fully connected layer for linear transformation to obtain an intermediate value of the risk assessment score, and the calculation formula of the intermediate value of the risk assessment score is as follows:

[0102] s = Wf feature +b

[0103] wherein W is the weight matrix of the fully connected layer, and b is the bias term.

[0104] In order to limit the value of the risk assessment score within a reasonable range, the intermediate value of the risk assessment score is mapped to the range of (0, 100) through mapping, and the calculation formula of the mapping process is as follows:

[0105]

[0106] wherein S scaled is the final risk assessment score, and its range is (0, 100).

[0107] According to the final risk assessment score, i.e., the anesthetic risk assessment guidance score value, the doctor can judge the potential risk of preoperative anesthesia, provide a reference for the doctor, and make the anesthetic risk assessment result more accurate.

[0108] For high-risk patients, the system will issue an alarm and prompt the doctor to conduct a more detailed review of the assessment content, or remind the doctor to strengthen targeted inquiry and obtain more patient condition information, and clearly display the anesthetic evaluation abnormal indicators of each patient in the form of a chart and a risk level, helping the doctor to make a quick decision.

[0109] Then, the doctor views the complete basic information, medical history information, photographed image information, and recorded video information of the patient on the evaluation data page through the management and auditing module of the APP doctor terminal, manually completes the anesthesia risk assessment, and can add remarks during the auditing process or edit and supplement the anesthesia risk assessment result that has been completed but not formally submitted, to ensure that the final anesthesia assessment report is accurate and reliable. For the anesthesia assessment report with the evaluation result of "pass", the doctor can perform electronic signature on the anesthesia informed consent form pushed by the module, upload the signed data to the third-party system CA electronic signature server for storage by calling the CA signature interface. If the evaluation result is "not pass", a notification of "need to supplement relevant examination and then re-visit, or need offline consultation" is given.

[0110] Later, the patient can view the anesthesia assessment report through the report preview module of the APP patient terminal, and the anesthesia assessment report includes patient basic information, medical history information, photographed image information, recorded video information, anesthesia risk assessment result, etc., and the display mode is a clear combination of text and images. The patient can also download or export the anesthesia assessment report as a PDF file for future reference or sharing with other medical institutions. If the anesthesia risk assessment result is "pass", the patient can perform electronic signature on the anesthesia informed consent form pushed by the evaluation result acquisition module, and upload the signed data to the third-party system CA electronic signature server for storage by calling the CA signature interface; if the anesthesia risk assessment result is "need to supplement relevant examination and then re-visit, or need offline consultation", the patient needs to complete relevant examination and then re-visit online or perform offline anesthesia assessment, and this evaluation is completed.

[0111] Finally, the doctor can check the completed anesthesia assessment report in the historical record module of the APP doctor terminal, and can retrieve the historical evaluation record of any patient at any time, and view the basic information, medical history information, photographed image information, recorded video information, anesthesia risk assessment result, etc. of the patient. The doctor can also quickly retrieve the required historical evaluation record according to the patient's name, evaluation time, operation type, etc. through the filtering and searching function of the historical record module, and recheck the evaluated result or supplement and modify the medical history when necessary.

[0112] The preoperative anesthesia assessment intelligent management system provided by the embodiment of the application can also perform offline anesthesia assessment. The main difference between the offline assessment process and the online assessment process is the system entry method, and the subsequent operations are basically the same, which are briefly described as follows:

[0113] When offline anesthesia assessment, the patient on the offline hospital end through the registration window / self-service machine end / hospital app port, identify the patient identity module after registration, identity recognition methods include card (such as, medical insurance card, ID card), scan code (such as, medical insurance two-dimensional code, clinic code). Through these identity recognition information, call the ESB interface provided by HIS to obtain the patient's detailed information (such as: patient HIS system unique ID, name, sex, age, etc.). Then, the patient completes the standardized form filling through the medical history collection module on the self-service machine, forming the final standardized form.

[0114] After that, the data automatic pulling and pushing function of the management audit module of the PC doctor end starts, and through the ESB interface provided by the HIS system, the patient registered for anesthesia assessment outpatient is grabbed and imported into the anesthesia assessment intelligent management system, reminding the doctor to handle as soon as possible.

[0115] Then, the intelligent analysis module of the PC doctor end starts, which gives a anesthesia risk assessment guidance score based on the pulled standardized form through intelligent analysis of the patient's patient basic information, medical history information, photographed image information, and recorded video information.

[0116] Then, the doctor calls the patient into the examination room according to the order of the call number through the call number system, and the doctor checks the patient's complete basic information, medical history information, photographed image information, recorded video information, supplements the medical history and vital signs, and manually completes the anesthesia risk assessment.

[0117] Then, the patient can view the anesthesia assessment report in the report preview module of the self-service machine, and can also download or export the anesthesia assessment report as a PDF file. If the anesthesia risk assessment result is "pass", go to the assessment result acquisition module to obtain the anesthesia informed consent form for electronic signature; if the anesthesia risk assessment result is "not pass", give the notice "need to supplement relevant examination and then re-visit" and complete the assessment.

[0118] Finally, the doctor can also check the completed anesthesia assessment report in the history record module of the PC doctor end, retrieve the historical assessment record of the evaluated patient at any time, and review or supplement and modify the history as necessary, and display the modification record.

[0119] In summary, the preoperative anesthesia evaluation intelligent management system proposed in the present application provides a comprehensive evaluation management tool for patients, doctors and hospital managers through the multi-platform integration of online Internet and offline hospital terminals. The system includes APP patient end, self-service machine end, PC doctor end and APP doctor end. Patients can perform self-evaluation through mobile devices or self-service terminals, and doctors can view and audit online in real time, effectively improving the efficiency of anesthesia evaluation. Further, the system ensures the standardization of the evaluation process through standardized forms and uses intelligent analysis modules to quantitatively analyze patient information, helping doctors reduce subjective judgments, make quick decisions, and improve the objectivity and accuracy of evaluation results. The intelligent guidance module set on the APP patient end and the self-service machine end provides animation and voice prompts for patients, helping them complete the evaluation smoothly, especially for the elderly and special patients, greatly improving the user experience. The system also connects the third-party system CA electronic signature server through the CA signature interface, truly realizing informationization and paperless management, and electronic signature confirmation ensures the security and compliance of data. The management audit module set on the PC doctor end and the APP doctor end has functions such as online communication, data synchronization, priority sorting, etc., making the evaluation process more efficient. In addition, the statistical analysis module of the PC doctor end also supports multi-dimensional data analysis and personalized report generation, providing support for clinical decision-making. Further, the system is also connected with the hospital's HIS system, realizing centralized management and real-time synchronization of patient information, avoiding the problem of data silos. As can be seen, the system solves the problems of low efficiency, incoherent data, strong subjectivity, and non-standard process in traditional anesthesia evaluation process through intelligent, standardized, information-based and multi-terminal interconnection, improving the efficiency, accuracy and patient experience of preoperative anesthesia evaluation.

[0120] Finally, it should be noted that although the embodiments of the present application have been described above, the present application is not limited to the specific embodiments and application fields described above, and the above specific embodiments are only illustrative and guiding, but not limiting. Those skilled in the art can make many forms under the guidance of the present specification without departing from the scope protected by the claims of the present application, and these all belong to the protection of the present application.

Claims

1. A preoperative anesthesia evaluation intelligent management system, characterized in that, The online internet end and the offline hospital end are included. The online internet end includes: An APP patient end for patients to perform preoperative self-evaluation and information interaction on a mobile terminal device; An APP doctor end for medical staff to view preoperative evaluation information of patients and give anesthesia risk assessment results on a mobile terminal device; The offline hospital end includes: A self-service machine end for providing self-service evaluation services and information interaction for offline patients; A PC doctor end for medical staff to view preoperative evaluation information of patients and give anesthesia risk assessment results in a hospital or workstation; The content of the preoperative evaluation information includes patient basic information, medical history information, photographed image information and recorded video information; and the anesthesia risk assessment result includes passing and needing offline consultation; The APP patient end and the APP doctor end are connected through an internet server, the self-service machine end and the PC doctor end are connected through a hospital server, and the internet server and the hospital server are connected.

2. The anesthesia evaluation intelligent management system according to claim 1, wherein The APP patient end includes the following functional modules: A medical history collection module for filling in preoperative evaluation information and uploading the filled preoperative evaluation information to an internet server; An evaluation result acquisition module for extracting anesthesia risk assessment results from anesthesia evaluation reports, reminding patients to confirm the anesthesia risk assessment results, pushing a anesthesia informed consent form when the anesthesia risk assessment result is "pass", and uploading signature data to an internet CA signature server after the patient performs electronic signature confirmation; A report preview module for obtaining anesthesia evaluation reports generated by the APP doctor end from the internet server and showing the anesthesia evaluation reports to the patient; the content of the anesthesia evaluation report includes patient basic information, medical history information, photographed image information, recorded video information, patient test reports and anesthesia risk assessment results; The APP doctor end includes the following functional modules: A management audit module for obtaining preoperative evaluation information of patients from the internet server, generating anesthesia evaluation reports based on anesthesia risk assessment results given by doctors and uploading the anesthesia evaluation reports to the internet server, pushing a anesthesia informed consent form when the anesthesia risk assessment result is "pass", and uploading signature data to an internet CA signature server after the medical staff perform electronic signature confirmation; A history record module for obtaining anesthesia evaluation reports from the internet server and archiving and storing the anesthesia evaluation reports; The self-service machine end includes all functional modules of the APP patient end, and the only difference is that the medical history collection module uploads the filled preoperative evaluation information to a hospital server; and the evaluation result acquisition module uploads signature data to a hospital CA signature server after the patient performs electronic signature confirmation; The report preview module obtains anesthesia evaluation reports generated by the APP doctor end from the hospital server; The self-service machine end further includes a patient identity recognition module for recognizing user identity and obtaining patient related information. The PC doctor end includes all the functional modules of the APP doctor end, the difference being that the management audit module obtains preoperative assessment information of a patient from an in-hospital server, and uploads an anesthesia assessment report to the in-hospital server, and uploads signature data to an in-hospital CA signature server after confirmation of electronic signature of medical staff; the historical record module is used to obtain the anesthesia assessment report from the in-hospital server.

3. The intelligent anesthesia assessment management system of claim 2, wherein, The APP doctor end and the PC doctor end further include an intelligent analysis module; the intelligent analysis module specifically includes: a data preprocessing unit, configured to automatically extract text, picture and video data in the preoperative assessment information, and pre-process the extracted data of different modalities respectively; a feature extraction unit, configured to perform feature extraction on the pre-processed text data and image data respectively, to obtain features of the data of different modalities; a feature fusion unit, configured to splice, reduce dimensions and pay attention to output of the features, to obtain fused features; a prediction output unit, configured to convert the fused features into a score value of anesthesia risk assessment guidance provided for reference of a doctor.

4. The intelligent anesthesia assessment management system of claim 2, wherein, The APP patient end and the self-service machine end further include an intelligent guidance module; the intelligent guidance module is used to provide animation guidance and voice prompts in the process of filling in the preoperative assessment information, to help a patient complete preoperative self-assessment.

5. The intelligent anesthesia assessment management system of claim 4, wherein, The intelligent guidance module is further used to convert recorded video information of the patient in the preoperative assessment information into quantitative data information according to corresponding actions of the patient according to the animation guidance and the voice prompts, by means of an artificial intelligence algorithm.

6. The intelligent anesthesia management system of claim 2, wherein, The management audit module further has at least one of the following functions: The management audit module is further used to collect and arrange the preoperative assessment information that has been submitted, and sort the preoperative assessment information according to submission time or patient priority; The management audit module further has a priority sorting function, and sorts the preoperative assessment information of the patient according to a surgery date or an emergency degree of a disease; The management audit module further has a doctor-patient online communication function, and supports online communication between a doctor and a patient.

7. The intelligent anesthesia management system of claim 2, wherein, The PC doctor end further includes a statistical analysis module; the statistical analysis module is used to perform multi-dimensional statistics and analysis on the assessment data, customize individualized reports, visually display data, and review historical data.

8. The intelligent anesthesia management system of claim 2, wherein, The PC doctor end further includes a system management module; the system management module is used for system parameter configuration, user permission management, system backup and recovery, log management and data security management.

9. The intelligent anesthesia management system of claim 2, wherein, The PC doctor end and the self-service machine end are further directly connected in communication, and are used to import relevant information of a patient registered for anesthesia assessment outpatient service into the anesthesia assessment intelligent management system; the PC doctor end is further connected in communication with at least one of an LIS system and a PACS system, and is used to obtain examination and inspection data of the patient.

10. A preoperative anesthesia evaluation intelligent analysis method, characterized in that, including: obtaining preoperative assessment information filled in by an online or in-hospital patient; the content of the preoperative assessment information includes patient basic information, medical history information, photographed image information and recorded video information; extracting text, picture and video data in the preoperative assessment information, and pre-processing the extracted data of different modalities respectively; performing feature extraction on the pre-processed text data and image data respectively, to obtain features of the data of different modalities; The features are spliced, dimensionally reduced, and attention outputted to obtain fused features; The fused features are converted into an anesthetic risk assessment guidance score value provided for reference by a doctor.

Images