In-hospital integrated platform system based on artificial intelligence

Abstract

The invention relates to the field of intelligent medical treatment, and discloses a hospital integrated platform system based on artificial intelligence, which comprises the steps of extracting current patient identification information, identifying a surgical patient in an anesthesia state in real time, and automatically associating a preoperative electronic medical record with an intraoperative anesthesia plan; based on an artificial intelligence semantic recognition model, performing image character extraction and semantic understanding on parameter contents set in real time in a respirator equipment screen; constructing a cross-system risk factor graph according to the semantic difference between the set parameters and the system identification, and performing risk linkage scoring on the operation behavior in the operation in real time; calling a parameter evolution path under the previous similar operation mode, and carrying out analogy modeling by combining the anesthesia depth of the current patient, the physiological data and the intraoperative stage; and when it is confirmed that strategy level abnormity exists in the setting, the early warning level and the display form are dynamically adjusted based on the operation equipment type, the setting personnel role and the load intensity of the current stage. The system has the advantage of improving the intelligent level of a hospital.

Description

Technical Field

[0001] The present invention relates to the field of smart medical care, and specifically to an in-hospital integrated platform system based on artificial intelligence. Background Art

[0002] During modern surgery, the ventilator is a crucial device for maintaining patient vital signs. This is especially true during general anesthesia, where spontaneous breathing is suppressed and the patient relies entirely on the ventilator for continuous and precise ventilation support. Multiple ventilator operating parameters, such as tidal volume, respiratory rate, and oxygen concentration, require dynamic adjustment based on intraoperative conditions and the patient's specific vital signs. Setting oxygen concentration is particularly critical; excessively high or low oxygen concentrations can lead to intraoperative complications. Currently, oxygen concentration settings are often manually input by anesthesiologists based on the type of surgery and patient condition. However, during rapid intraoperative patient transitions or shift handovers, incomplete information transfer or oversight can easily lead to misaligned parameters. Especially during emergency situations, medical staff often prioritize life support and pharmacological interventions, neglecting to review device settings. However, existing medical AI systems primarily focus on diagnostic and treatment decision support or image recognition, lacking the ability to perceive and address specific details of intraoperative device use. Currently, there is no systematic solution that can leverage AI to implement real-time interventions at the detailed level of device parameter settings, integrating data from upstream and downstream systems. Therefore, it is necessary to design an artificial intelligence-based in-hospital integrated platform system to improve the intelligence level of the hospital. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention provides an in-hospital integrated platform system based on artificial intelligence, which has the advantage of improving the intelligence level of the hospital and solves the problems in the above-mentioned background technology.

[0004] To achieve the above-mentioned purpose of improving the hospital's intelligence level, the present invention provides the following technical solutions: an artificial intelligence-based in-hospital integrated platform system, comprising:

[0005] Tag matching module: This module extracts the current patient identification information, identifies the surgical patient under anesthesia in real time, automatically links the preoperative electronic medical record with the intraoperative anesthesia plan, and determines whether there is a need to set the oxygen concentration. If so, it enters the parameter semantic parsing module.

[0006] Semantic Parsing Module: Based on an artificial intelligence semantic recognition model, it extracts image text and semantically understands the parameters set in real time on the ventilator device screen. Combined with preoperative contraindications and recommended oxygen concentration ranges in the electronic medical record, it determines whether there is a semantic deviation in the set value. If so, it enters the instruction risk linkage module.

[0007] Risk linkage module: Based on the semantic differences between the set parameters and the system recognition, a cross-system risk factor map is constructed to perform risk linkage scoring on intraoperative operation behaviors in real time to determine whether AI intervention conditions are triggered. If so, the intraoperative strategy comparison module is entered;

[0008] Strategy comparison module: retrieves the parameter evolution path of similar surgical procedures in the past, combines the current patient's anesthesia depth, physiological data and intraoperative stage to perform analog modeling, and determines whether there is a strategy deviation in the current set behavior. If so, it enters the early warning trigger feedback module;

[0009] Early warning feedback module: When it is confirmed that there is a policy-level anomaly in the setting, the early warning level and display format are dynamically adjusted based on the operating equipment type, the set personnel role and the current load intensity, and graphical intervention suggestions are generated.

[0010] Preferably, the process of determining whether there is an oxygen concentration setting requirement is as follows:

[0011] Extract the binding information between the intraoperative equipment and the patient, and read the patient's current anesthesia status identifier in the anesthesia information system;

[0012] Determine whether the patient's current surgical stage involves spontaneous respiratory suppression or gas management requirements;

[0013] Combined with the preoperative anesthesia plan and surgical requirements, compare the recommended oxygen concentration range with the current ventilator oxygen concentration setting to see if they are synchronized;

[0014] If the current set value is missing or deviates from the recommended range by more than the set ratio threshold, it is determined that there is an oxygen concentration setting requirement.

[0015] Preferably, the process of image text extraction and semantic understanding of the parameter content set in real time on the ventilator device screen is:

[0016] Obtain the image stream of the ventilator panel area and perform preprocessing operations on the image area, including edge sharpening, color enhancement and distortion correction;

[0017] Using the trained parameter positioning model, identify the oxygen concentration parameter label and setting value block in the display area;

[0018] Extract the set value and convert it into a semantic label based on the device manufacturer's parameter layout dictionary.

[0019] Preferably, the process of determining whether the set value has a semantic deviation is:

[0020] Match the extracted oxygen concentration setting value with the recommended oxygen concentration range in the patient's preoperative electronic medical record to see if the current setting value falls within the recommended range;

[0021] Extract high-risk factor markers for patients and construct a clinical context risk factor set;

[0022] Using a semantic bias identification model based on a knowledge graph, the set value is integrated with the clinical context to calculate a bias risk score;

[0023] When the bias risk score is greater than the policy tolerance, it is determined that semantic bias exists;

[0024] When the bias risk score is less than or equal to the policy tolerance, it is determined that there is no semantic bias.

[0025] Preferably, the process of constructing a cross-system risk factor map is:

[0026] Aggregate multi-source data related to patient status from ventilators, anesthesia information systems, electronic medical record systems, and monitoring systems, including setting parameters, historical oxygen usage curves, intraoperative monitoring indicators, and handover operation records;

[0027] Standardize the data and map it to the risk factor graph template, and use graph embedding technology to build cross-system node connections and dependency paths;

[0028] Mark potential abnormal paths and output a cross-system risk factor map.

[0029] Preferably, the process of real-time risk linkage scoring of intraoperative operation behavior is as follows:

[0030] Extract the system context status at the time when the current set behavior occurs, including the operator role, set time point, physiological parameter status curve and equipment load status;

[0031] Based on the correlation paths in the cross-system risk factor map, identify whether the operation behavior triggers a high-weight risk link;

[0032] The timing risk scoring model is used to perform weighted calculation on the triggering time and strength of each link to form a risk linkage score for the operation behavior.

[0033] Preferably, the process of determining whether the AI ​​intervention condition is triggered is:

[0034] Fusion analysis of the risk linkage score of the operation behavior and the patient status stability index;

[0035] Dynamically adjust the risk level by combining the emergency level of the anesthesia stage with the current operation intensity of the equipment operators;

[0036] When the combined score of the fusion risk index and the actual operation intensity exceeds the dynamic intervention threshold, the AI ​​intervention process is triggered;

[0037] When the combined score of the fusion risk index and the actual operation intensity is less than or equal to the dynamic intervention threshold, the AI ​​intervention process will not be triggered.

[0038] Preferably, the analog modeling process is performed by combining the current patient anesthesia depth, physiological data and intraoperative stage as follows:

[0039] Retrieve historical cases with similar anesthesia methods, surgical procedures, and underlying patient disease characteristics from the structured database, and select sample sets with similarity to the current cases greater than a threshold.

[0040] Extract the oxygen concentration setting change trajectory, intervention response time, and final postoperative recovery data from these samples to form a strategy evolution template;

[0041] Combining the patient's current BIS anesthesia depth, hemodynamic status, and intraoperative procedure labels, an analogy modeling algorithm is used to generate a predicted risk distribution for the current strategy in a historical context.

[0042] Output the degree of match between the current set behavior and the risk set path in the analogy template.

[0043] Preferably, the process of determining whether there is a policy deviation in the current set behavior is:

[0044] Compare the current set oxygen concentration value with the optimal trajectory range in the analog model to determine the amplitude and time dimension of the deviation range;

[0045] Combined with the patient's real-time vital sign change trends when the set behavior occurs, the subsequent physiological chain reactions triggered can be evaluated;

[0046] Based on the strategy deviation scoring function, the deviation degree of the current setting behavior from the analog historical model is quantified, and corrections are made based on the historical operating habits of the setting personnel and the error model;

[0047] When the strategy deviation score exceeds the warning value predicted by the system's self-learning model, it indicates that there is a strategy deviation in the current set behavior.

[0048] Preferably, the process of generating graphical intervention suggestions is:

[0049] Automatically selects warning presentation templates based on the current deviation type, surgical stage, and device type, distinguishing color levels, shape identifiers, and information display levels;

[0050] By integrating specific indicators of strategy deviation with the system judgment logic, a graphical suggestion box containing parameter value comparison, risk prediction trend chart and recommended adjustment range is generated on the interface to generate graphical intervention suggestions.

[0051] Compared with the existing technology, the present invention provides an in-hospital integrated platform system based on artificial intelligence, which has the following beneficial effects:

[0052] The present invention automatically associates the patient's identity with key preoperative information through a label matching module, accurately identifying whether the setting behavior needs attention; with the help of a semantic analysis module, image text extraction and clinical semantic judgment of the device screen setting parameters are realized, significantly improving the real-time and accuracy of parameter identification; the risk linkage module integrates cross-system data and knowledge graph technology to score operational behaviors and potential risk paths, and has dynamic response capabilities; the strategy comparison module uses analogy modeling technology to compare and analyze the current setting behavior with the historical optimal strategy trajectory, and strengthens the strategy-level deviation identification capability; the early warning feedback module combines the operator's role, equipment load and intraoperative stage to intelligently adjust the intervention level and the way the suggestion is presented, thereby improving clinical acceptability and interactive friendliness. The system as a whole has improved the level of intelligence in anesthesia management, reduced intraoperative risks caused by setting errors, and has good clinical practicality and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram of the system of the present invention. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] Example 1: Please refer to Figure 1 As shown, an artificial intelligence-based hospital integration platform system according to an embodiment of the present invention includes:

[0056] Label matching module: Extracts the current patient identification information, identifies surgical patients under anesthesia in real time, automatically associates preoperative electronic medical records with intraoperative anesthesia plans, and determines whether there is a need to set oxygen concentration. If so, it enters the parameter semantic parsing module.

[0057] The process of determining whether there is an oxygen concentration setting requirement in the tag matching module is as follows:

[0058] Extract the binding information between the intraoperative equipment and the patient, and read the patient's current anesthesia status identification in the anesthesia information system; after the patient enters the operating room, the system automatically completes the patient's identity identification through surgical scheduling information, wristband identification code, electronic medical record number, etc., and uniquely binds with key intraoperative equipment such as anesthesia machine or ventilator. The equipment uploads its equipment number, equipment type and network address information through its built-in communication interface. The system establishes a mapping relationship between the equipment information and the patient's unique identity and records the binding time point. The system accesses the anesthesia information system through the interface protocol and extracts the current patient's anesthesia stage information. The anesthesia stage usually includes anesthesia preparation period, induction period, maintenance period, awakening period and recovery period. Each stage is manually marked by the anesthesiologist or automatically identified by the system based on events such as drug injection time and respiratory status, and the current status identification code is updated;

[0059] Determine whether the patient's current surgical stage involves spontaneous respiratory suppression or gas management requirements; based on the patient's anesthesia stage, combined with the type of surgical procedure, ventilation method, and medication use, determine whether the current stage is a critical stage requiring precise management of inspired oxygen concentration. The judgment process includes the following dimensions:

[0060] Anesthesia stage judgment: If the current stage is the induction phase or maintenance phase, and the patient is monitored to have used muscle relaxants or sedatives, the system will preliminarily determine that the patient is in a state of suppressed spontaneous breathing;

[0061] Auxiliary judgment indicators: If the patient has endotracheal intubation, laryngeal mask insertion, or the breathing mode is fully controlled ventilation, it is further confirmed that the patient is unable to independently adjust the composition of the inhaled gas;

[0062] Identifying surgical procedure requirements: If the surgical type is thoracoscopic surgery, one-lung ventilation, prone spinal surgery, or other procedures involving lung isolation and high ventilation requirements, the system defaults to requiring precise control of oxygen concentration. The surgical procedure information is provided by the ProcedureType field in the preoperative plan. The system determines whether the field contains the GasControlRequired flag.

[0063] In summary, if any of the judgment conditions is met, the system will consider that the setting state of the inhaled oxygen concentration needs to be paid attention to at the current stage;

[0064] Combine the preoperative anesthesia plan with the surgical requirements and compare the recommended oxygen concentration range with the current ventilator oxygen concentration setting to see if they are synchronized. Read the preoperative anesthesia plan form, which usually includes the recommended inspired oxygen concentration range based on the patient's underlying disease, surgical complexity, and intraoperative ventilation goals. For example, for patients without underlying respiratory diseases, the recommended range is 30%–40%. If chronic obstructive pulmonary disease is present or low tidal volume ventilation is planned during surgery, the recommended range may be adjusted to 35%–50% or higher.

[0065] If the current set value is missing or deviates from the recommended range by more than the set ratio threshold, it is determined that there is an oxygen concentration setting requirement.

[0066] Semantic parsing module: Based on the artificial intelligence semantic recognition model, it extracts image text and semantically understands the parameter content set in real time on the ventilator equipment screen. Combined with the preoperative contraindications and recommended oxygen concentration range in the electronic medical record, it determines whether there is a semantic deviation in the set value. If so, it enters the instruction risk linkage module.

[0067] The semantic parsing module performs image text extraction and semantic understanding on the parameter content set in real time on the ventilator device screen as follows:

[0068] Obtain the image stream of the ventilator panel area and perform preprocessing operations on the image area, including edge sharpening, color enhancement and distortion correction;

[0069] Edge sharpening: Use Sobel or Laplacian operators to enhance the edges of characters on the display screen, improving the clarity of the recognition boundaries of numbers and units;

[0070] Color enhancement: Apply histogram equalization or LAB color space adjustment algorithms to improve screen brightness and contrast, and compensate for screen reflections and low light effects;

[0071] Distortion correction: If the image has perspective distortion caused by perspective offset or wide-angle lens, use affine transformation or OpenCV's perspective transformation function to correct it;

[0072] The trained parameter localization model is used to identify oxygen concentration parameter labels and set value blocks in the display area. A lightweight object detection network is used to locate the parameter area. The parameter localization model accepts preprocessed screen images as input and outputs labeled bounding box coordinates. The model training samples cover ventilator interface images from multiple manufacturers, and the annotation content includes: oxygen concentration label text and set value display area;

[0073] Extract the set value and convert it into a semantic label based on the device manufacturer's parameter layout dictionary.

[0074] The process of judging whether the set value has semantic deviation in the semantic parsing module is as follows:

[0075] The extracted oxygen concentration setting value is matched with the recommended oxygen concentration range in the patient's preoperative electronic medical record to compare whether the current setting value falls within the recommended range; the current set oxygen concentration value is obtained in real time from the clinical monitoring system or anesthesia machine interface, expressed in digital form, such as 40%, 60%, etc.; the patient's preoperative electronic medical record (EMR) is parsed to extract the recommended oxygen concentration range, which is given by the anesthesiology department or attending physician in the preoperative assessment based on the patient's underlying disease (such as COPD, asthma, heart failure) and other risk factors. For example, the recommended oxygen concentration is [30%–50%]; a matching algorithm is used to judge the interval between the setting value and the recommended range. If the setting value falls within the range, it is preliminarily judged as "no deviation"; if it falls outside the range, the next step is to determine whether it is a semantic deviation (i.e., whether there is a risk difference between the current setting value and the recommended value due to clinical semantic factors);

[0076] Extract high-risk factor markers from patients and construct a clinical contextual risk factor set; extract high-risk factor information from EMRs, preoperative assessment forms, ICD diagnoses, and laboratory test results, including but not limited to: underlying respiratory diseases (e.g., COPD, pulmonary fibrosis); cardiovascular diseases (e.g., coronary heart disease, heart failure); advanced age (e.g., >75 years) or low weight (BMI <18); and surgical characteristics (e.g., thoracotomy, neurosurgery); perform structured processing on the extracted risk factors and convert them into labels or vectors that can be used as model input;

[0077] Using a semantic deviation recognition model based on knowledge graph, the set value and clinical background are integrated and calculated to obtain a deviation risk score; a medical knowledge graph is introduced (such as the relationship between disease-treatment-risk-physiological indicators), and the nodes in the graph cover disease names, recommended treatment parameters, safety ranges, and concurrent risks; the set value, recommended interval, and risk factor set are mapped to the corresponding nodes of the knowledge graph to establish semantic context links, for example: oxygen concentration = 60% → high oxygen inhalation risk; COPD + high oxygen concentration → potential Increased retention risk; using graph neural networks or graph attention mechanisms to aggregate the current set value and the patient graph context to generate a bias risk score, usually a continuous value between 0 and 1, with higher scores indicating greater risk of semantic bias;

[0078] When the bias risk score is greater than the policy tolerance, it is determined that semantic bias exists;

[0079] When the bias risk score is less than or equal to the policy tolerance, it is determined that there is no semantic bias.

[0080] Risk linkage module: Based on the set parameters and the semantic differences identified by the system, a cross-system risk factor map is constructed, and risk linkage scoring of intraoperative operation behaviors is performed in real time to determine whether AI intervention conditions are triggered. If so, the intraoperative strategy comparison module is entered.

[0081] The process of constructing a cross-system risk factor map in the risk linkage module is as follows:

[0082] Aggregate multi-source data related to patient status from ventilators, anesthesia information systems, electronic medical record systems, and monitoring systems, including setting parameters, historical oxygen usage curves, intraoperative monitoring indicators, and handover operation records;

[0083] Standardize the data and map it to the risk factor graph template, and use graph embedding technology to build cross-system node connections and dependency paths;

[0084] Mark potential abnormal paths, such as conflicts between preoperative oxygen contraindications and intraoperative settings, inconsistencies between equipment settings and patient vital sign trends, etc., and output a cross-system risk factor map.

[0085] The process of real-time risk linkage scoring of intraoperative operation behaviors in the risk linkage module is as follows:

[0086] Extract the system context status at the time when the current set behavior occurs, including the operator role, set time point, physiological parameter status curve and equipment load status;

[0087] Based on the correlation path in the cross-system risk factor map, identify whether the operation behavior triggers a high-weight risk link; map the current set behavior to the corresponding node in the risk factor map, such as setting =80%→node: hyperoxia setting; use the graph index mechanism to quickly retrieve the upstream and downstream association structure of the node in the graph; traverse the risk path directly or indirectly connected to the operation behavior node to identify whether it forms a closed chain (such as COPD→hyperoxia → Each link has a preset weight (from clinical knowledge graph, data statistics or expert scoring), such as 0.85 for the hyperoxia link and 0.75 for the low-weight intraoperative anesthesia fluctuation link; for the path triggered by the operation behavior, it is determined whether the complete triggering conditions are met (for example, all key condition nodes on the graph chain are activated), otherwise it is not recorded as a risk linkage; Example: Only when "COPD" + "hyperoxia" ”+“ The hyperoxia risk chain is activated only when the three conditions of "decline" and "decrease" co-occur. The same operation behavior may trigger multiple links. The system detects all related risk chains in parallel and prepares them for input into the temporal risk scoring model.

[0088] The timing risk scoring model is used to perform weighted calculation on the triggering time and strength of each link to form a risk linkage score for the operation behavior.

[0089] The process of determining whether to trigger AI intervention conditions in the risk linkage module is as follows:

[0090] Fusion analysis of the risk linkage score of the operation behavior and the patient status stability index;

[0091] Dynamically adjust the risk level by combining the emergency level of the anesthesia stage with the current operation intensity of the equipment operators;

[0092] When the combined score of the fusion risk index and the actual operation intensity exceeds the dynamic intervention threshold, the AI ​​intervention process is triggered;

[0093] When the combined score of the fusion risk index and the actual operation intensity is less than or equal to the dynamic intervention threshold, the AI ​​intervention process will not be triggered.

[0094] Strategy comparison module: retrieves the parameter evolution path of similar surgical procedures in the past, combines the current patient's anesthesia depth, physiological data and intraoperative stage to perform analog modeling, and determines whether there is a strategy deviation in the current set behavior. If so, enter the early warning trigger feedback module.

[0095] The analog modeling process in the strategy comparison module is based on the current patient anesthesia depth, physiological data and intraoperative stage:

[0096] Retrieve historical cases with similar anesthesia methods, surgical procedures, and underlying patient disease characteristics from the structured database, and select sample sets with similarity to the current cases greater than a threshold.

[0097] Extract the oxygen concentration setting change trajectory, intervention response time, and final postoperative recovery data from these samples to form a strategy evolution template; ) set value time series reconstruction to form a "setting trajectory diagram"; capture the specific moment, amplitude and direction (increase or decrease) of each setting adjustment; at the same time, the depth of anesthesia (BIS value), 、 Align the synchronous physiological index curves; mark the target index after each setting adjustment (such as ) The time required for a noticeable response (such as an increase or stabilization) to occur; for example: After rising from 50% to 70%, The time required to increase from 89% to 95% was 180 seconds; marker adjustment-response pairing; extraction of postoperative PACU recovery score, whether to be transferred to ICU, postoperative complications (such as retention, atelectasis) and other information; label the strategy trajectory as "low risk," "caution," "high risk," and other levels based on the results; classify and cluster the trajectories and results of multiple similar samples to form a strategy evolution template:

[0098] Template A: Steady-state oxygen strategy → Improved stability → good recovery after surgery;

[0099] Template B: Sudden oxygen concentration increase + high BIS fluctuation → Slow recovery → moderate delay in recovery;

[0100] Template C: Frequently adjust oxygen concentration → Severe fluctuations → increased ICU admission rate;

[0101] Combined with the current patient's BIS anesthesia depth, hemodynamic status and intraoperative operation labels, the analog modeling algorithm is used to generate the predicted risk distribution of the current setting strategy in the historical context; capture the current intraoperative setting behavior and its corresponding time point; synchronously extract the BIS value before and after the time point (such as the change trend from 68 to 58), HR / MAP curve, Changes, etc.; Synchronously obtain the set behavior tags (such as " 20% increase", operator authority level, occurrence stage, etc.); use a combination of K-nearest neighbor + probability model or deep analogy network (such as Siamese Network); match the current strategy characteristics with the trajectory in the strategy template; calculate the probability distribution of the risk level corresponding to the strategy in historical samples; output the risk probability distribution; and output the degree of match between the current set behavior and the risk setting path in the analogy template.

[0102] The process of judging whether there is a strategy deviation in the current setting behavior in the strategy comparison module is as follows:

[0103] Compare the current set oxygen concentration value with the optimal trajectory range in the analog modeling to determine the amplitude and time dimension of the deviation range; capture the current Set value and time point (for example: 35 minutes into the operation, From 60% to 80%); form a set behavior node with a timestamp and set amplitude label; call the optimal trajectory template previously generated by aggregating similar historical cases, which contains the ideal range of oxygen concentration settings; compare whether the current set value falls within the optimal trajectory tolerance interval at that time point; calculate the offset amplitude: Dm= ;Calculate the offset time window. If it is not set within the recommended time period, the time will be marked as misaligned;

[0104] Combined with the patient's real-time vital sign changes when the set behavior occurs, the subsequent physiological chain reaction triggered is evaluated; key parameters 3-5 minutes before and after the set behavior occur are extracted: 、 , MAP, HR, BIS; convert data into trend lines (up / down / fluctuation range) and derived indicators (such as rate of change, coefficient of variation); identify abnormal reaction chains after set behaviors, such as: A sharp rise → Decline → HR instability; Adjust → BIS fluctuates dramatically; compare with known physiological response pathways in the atlas to see if abnormal response tags are triggered; set impact factors and weights for each link to generate a cumulative score;

[0105] Based on the strategy deviation scoring function, the deviation degree of the current setting behavior from the analog historical model is quantified, and corrections are made based on the historical operating habits of the setting personnel and the error model;

[0106] When the strategy deviation score exceeds the warning value predicted by the system's self-learning model, it indicates that there is a strategy deviation in the current set behavior.

[0107] Early warning feedback module: When it is confirmed that there is a policy-level anomaly in the setting, the early warning level and display format are dynamically adjusted based on the operating equipment type, the set personnel role and the current load intensity, and graphical intervention suggestions are generated.

[0108] The process of generating graphical intervention suggestions in the early warning feedback module is as follows:

[0109] Automatically selects warning presentation templates based on the current deviation type, surgical stage, and device type, distinguishing color levels, shape identifiers, and information display levels;

[0110] By integrating specific indicators of strategy deviation with the system judgment logic, a graphical suggestion box containing parameter value comparison, risk prediction trend chart and recommended adjustment range is generated on the interface to generate graphical intervention suggestions.

[0111] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0112] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An artificial intelligence-based in-hospital integrated platform system, characterized by: include: Tag matching module: This module extracts the current patient identification information, identifies the surgical patient under anesthesia in real time, automatically links the preoperative electronic medical record with the intraoperative anesthesia plan, and determines whether there is a need to set the oxygen concentration. If so, it enters the parameter semantic parsing module. Semantic Parsing Module: Based on an artificial intelligence semantic recognition model, it extracts image text and semantically understands the parameters set in real time on the ventilator device screen. Combined with preoperative contraindications and recommended oxygen concentration ranges in the electronic medical record, it determines whether there is a semantic deviation in the set value. If so, it enters the instruction risk linkage module. Risk linkage module: Based on the semantic differences between the set parameters and the system recognition, a cross-system risk factor map is constructed to perform risk linkage scoring on intraoperative operation behaviors in real time to determine whether AI intervention conditions are triggered. If so, the intraoperative strategy comparison module is entered; Strategy comparison module: retrieves the parameter evolution path of similar surgical procedures in the past, combines the current patient's anesthesia depth, physiological data and intraoperative stage to perform analog modeling, and determines whether there is a strategy deviation in the current set behavior. If so, it enters the early warning trigger feedback module; Early warning feedback module: When it is confirmed that there is a policy-level anomaly in the setting, the early warning level and display format are dynamically adjusted based on the operating equipment type, the set personnel role and the current load intensity, and graphical intervention suggestions are generated.

2. The artificial intelligence-based hospital integrated platform system according to claim 1, characterized in that: The process of determining whether there is a need to set the oxygen concentration is as follows: Extract the binding information between the intraoperative equipment and the patient, and read the patient's current anesthesia status identifier in the anesthesia information system; Determine whether the patient's current surgical stage involves spontaneous respiratory suppression or gas management requirements; Combined with the preoperative anesthesia plan and surgical requirements, compare the recommended oxygen concentration range with the current ventilator oxygen concentration setting to see if they are synchronized; If the current set value is missing or deviates from the recommended range by more than the set ratio threshold, it is determined that there is an oxygen concentration setting requirement.

3. The artificial intelligence-based hospital integrated platform system according to claim 2, characterized in that: The process of image text extraction and semantic understanding of the real-time parameter settings on the ventilator device screen is as follows: Obtain the image stream of the ventilator panel area and perform preprocessing operations on the image area, including edge sharpening, color enhancement and distortion correction; Using the trained parameter positioning model, identify the oxygen concentration parameter label and setting value block in the display area; Extract the set value and convert it into a semantic label based on the device manufacturer's parameter layout dictionary.

4. The artificial intelligence-based hospital integrated platform system according to claim 3, characterized in that: The process of judging whether the set value has semantic deviation is as follows: Match the extracted oxygen concentration setting value with the recommended oxygen concentration range in the patient's preoperative electronic medical record to see if the current setting value falls within the recommended range; Extract high-risk factor markers for patients and construct a clinical context risk factor set; Using a semantic bias identification model based on a knowledge graph, the set value is integrated with the clinical context to calculate a bias risk score; When the bias risk score is greater than the policy tolerance, it is determined that semantic bias exists; When the bias risk score is less than or equal to the policy tolerance, it is determined that there is no semantic bias.

5. The artificial intelligence-based hospital integrated platform system according to claim 4, characterized in that: The process of constructing a cross-system risk factor map is as follows: Aggregate multi-source data related to patient status from ventilators, anesthesia information systems, electronic medical record systems, and monitoring systems, including setting parameters, historical oxygen usage curves, intraoperative monitoring indicators, and handover operation records; Standardize the data and map it to the risk factor graph template, and use graph embedding technology to build cross-system node connections and dependency paths; Mark potential abnormal paths and output a cross-system risk factor map.

6. The artificial intelligence-based hospital integrated platform system according to claim 5, characterized in that: The process of real-time risk linkage scoring of intraoperative operation behaviors is as follows: Extract the system context status at the time when the current set behavior occurs, including the operator role, set time point, physiological parameter status curve and equipment load status; Based on the correlation paths in the cross-system risk factor map, identify whether the operation behavior triggers a high-weight risk link; The timing risk scoring model is used to perform weighted calculation on the triggering time and strength of each link to form a risk linkage score for the operation behavior.

7. The artificial intelligence-based hospital integrated platform system according to claim 6, characterized in that: The process of determining whether AI intervention conditions are triggered is as follows: Fusion analysis of the risk linkage score of the operation behavior and the patient status stability index; Dynamically adjust the risk level by combining the emergency level of the anesthesia stage with the current operation intensity of the equipment operators; When the combined score of the fusion risk index and the actual operation intensity exceeds the dynamic intervention threshold, the AI ​​intervention process is triggered; When the combined score of the fusion risk index and the actual operation intensity is less than or equal to the dynamic intervention threshold, the AI ​​intervention process will not be triggered.

8. The artificial intelligence-based hospital integrated platform system according to claim 7, characterized in that: The analog modeling process based on the current patient anesthesia depth, physiological data and intraoperative stage is as follows: Retrieve historical cases with similar anesthesia methods, surgical procedures, and underlying patient disease characteristics from the structured database, and select sample sets with similarity to the current cases greater than a threshold. Extract the oxygen concentration setting change trajectory, intervention response time, and final postoperative recovery data from these samples to form a strategy evolution template; Combining the patient's current BIS anesthesia depth, hemodynamic status, and intraoperative procedure labels, an analogy modeling algorithm is used to generate a predicted risk distribution for the current strategy in a historical context. Output the degree of match between the current set behavior and the risk set path in the analogy template.

9. The artificial intelligence-based hospital integrated platform system according to claim 8, characterized in that: The process of judging whether there is a policy deviation in the current setting behavior is as follows: Compare the current set oxygen concentration value with the optimal trajectory range in the analog model to determine the amplitude and time dimension of the deviation range; Combined with the patient's real-time vital sign change trends when the set behavior occurs, the subsequent physiological chain reactions triggered can be evaluated; Based on the strategy deviation scoring function, the deviation degree of the current setting behavior from the analog historical model is quantified, and corrections are made based on the historical operating habits of the setting personnel and the error model; When the strategy deviation score exceeds the warning value predicted by the system's self-learning model, it indicates that there is a strategy deviation in the current set behavior.

10. The artificial intelligence-based hospital integrated platform system according to claim 9, characterized in that: The process of generating graphical intervention recommendations is as follows: Automatically selects warning presentation templates based on the current deviation type, surgical stage, and device type, distinguishing color levels, shape identifiers, and information display levels; By integrating specific indicators of strategy deviation with the system judgment logic, a graphical suggestion box containing parameter value comparison, risk prediction trend chart and recommended adjustment range is generated on the interface to generate graphical intervention suggestions.

Images