Orthopedic patient intraoperative nursing comprehensive management system based on artificial intelligence

Abstract

The present application relates to the field of medical information technology, in particular to a comprehensive management system for intraoperative nursing of orthopedic patients based on artificial intelligence, comprising event analysis, rule extraction, state snapshot, strategy decision, sequence control and iterative update modules. The system analyzes the monitoring signals of heterogeneous terminals, dynamically calls key sign indicators and warning thresholds from the knowledge graph according to the operation stage, and generates a phased sign state snapshot. By comparing the sign data with the threshold, the reasoning engine is activated, the composite nursing instructions containing multiple collaborative measures are output, the virtual nursing task execution sequence is built and tracked in the system, and then the snapshot is iteratively updated according to the task completion state and new perception events, forming a closed-loop management. The scheme realizes the dynamic adaptation of intraoperative monitoring thresholds to the operation process, and integrates discrete nursing operations into a structured task process that can be tracked and evaluated.

Description

Technical Field

[0001] This invention relates to the field of medical information technology, and in particular to an artificial intelligence-based comprehensive management system for intraoperative nursing care of orthopedic patients. Background Technology

[0002] Currently, vital sign monitoring during orthopedic surgery generally relies on independent monitoring devices. These devices are typically preset with fixed alarm thresholds based on general medical standards or the experience of anesthesiologists. Nursing staff need to simultaneously monitor multiple device screens, receive discrete audible and visual alarms when indicators exceed the thresholds, and make judgments and take actions based on their personal experience. This is a passive, reactive monitoring model, lacking systematic coordination and recording between nursing responses.

[0003] This conventional technical approach has its flaws. Surgery is a dynamic and continuous process, with varying degrees of physiological disturbance to the patient at different stages. Fixed monitoring thresholds cannot adapt to these stage-specific changes, easily generating meaningless alarms that interfere with medical care in low-risk stages, while potentially delaying warnings in high-risk stages due to insufficient sensitivity of the thresholds. Furthermore, discrete alarms and manual interventions fragment nursing actions, making it difficult to systematically track and verify the logical sequence and effectiveness of various measures, thus failing to form a closed-loop evaluation of the effectiveness of nursing intervention strategies.

[0004] The challenge lies in addressing two key issues: how to dynamically adapt physiological monitoring standards to different stages of the surgical procedure, and how to integrate discrete nursing responses into a traceable and assessable sequence of collaborative tasks. The former aims to shift from static alarms to dynamic intelligent early warning, while the latter aims to elevate management from manual, discrete interventions to a systematic, closed-loop system. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an artificial intelligence-based comprehensive management system for intraoperative nursing care of orthopedic patients.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an artificial intelligence-based comprehensive management system for intraoperative nursing care of orthopedic patients, comprising: The event parsing module acquires raw monitoring signals about the patient's vital signs from multiple heterogeneous sensing terminals within the surgical environment, and parses the raw monitoring signals into discrete vital sign sensing events with a unified data structure. The rule extraction module extracts a set of key signs associated with different stages of orthopedic surgery from a preset intraoperative nursing knowledge graph, as well as the dynamic warning threshold of each key sign in the set of key signs. The status snapshot module maps the discrete vital sign perception events to the set of key vital sign indicators to generate a staged vital sign status snapshot corresponding to the current stage type of the orthopedic surgery. The strategy decision-making module, based on the comparison results of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold, activates the preset nursing strategy reasoning engine and outputs composite nursing instructions. The sequence control module, based on the composite nursing instructions, constructs a virtual nursing task execution sequence within the AI-based orthopedic patient intraoperative nursing comprehensive management system, and tracks the completion status of each task in the virtual nursing task execution sequence in real time. The iterative update module dynamically aggregates new discrete vital sign perception events from the heterogeneous sensing terminal based on the completion status of the virtual nursing task execution sequence, so as to iteratively update the phased vital sign status snapshot.

[0007] Preferably, the step of parsing the original monitoring signal into discrete vital sign perception events with a unified data structure includes: The system receives an analog voltage waveform sequence from an electrocardiogram monitor, an inflation / deflation pressure data stream from a non-invasive blood pressure monitor, and an impedance change value from a body temperature sensor. The analog voltage waveform sequence, the inflation / deflation pressure data stream, and the impedance change value together constitute the original monitoring signal. The analog voltage waveform sequence is subjected to heartbeat segmentation based on periodic features, the inflation and deflation pressure data stream is subjected to blood pressure cycle identification based on pressure inflection points, and the impedance change value is subjected to mean filtering based on time windows to obtain a set of heartbeat segments, independent blood pressure measurement cycles, and filtered body temperature time series data, respectively. For each heartbeat segment in the aforementioned heartbeat segment set, the RR interval, QRS complex amplitude and width are calculated to generate a standardized electrocardiogram event; For each independent blood pressure measurement cycle, the peak systolic blood pressure, the trough diastolic blood pressure, and the mean arterial pressure are extracted to generate standardized blood pressure events. The filtered body temperature time-series data is sampled at fixed time intervals to generate standardized body temperature events; The standardized electrocardiogram event, the standardized blood pressure event, and the standardized body temperature event are uniformly encapsulated into a discrete vital sign perception event containing a timestamp, vital sign type code, numerical value, and source identifier.

[0008] Preferably, the step of extracting a set of key vital signs associated with different stages of orthopedic surgery and a dynamic warning threshold for each key vital sign in the set of key vital signs from a preset intraoperative nursing knowledge graph includes: Receive the current surgical stage code sent by the surgical anesthesia information system. The surgical stage code includes the anesthesia induction stage, surgical incision stage, key instrument operation stage, and wound closure stage. Query the nodes associated with the anesthesia induction phase in the preset intraoperative nursing knowledge graph, obtain heart rate variability and blood pressure drop rate as the first phase key sign indicator set, and extract the dynamic warning threshold of the first phase key sign indicator set related to age and baseline blood pressure. Query the nodes associated with the surgical incision stage in the preset intraoperative nursing knowledge graph, obtain the instantaneous increase in heart rate and blood pressure as the second stage key sign indicator set, and extract the dynamic warning threshold of the second stage key sign indicator set related to the patient's pain threshold. Query the nodes associated with the key instrument operation stage in the preset intraoperative nursing knowledge graph, obtain the core body temperature change trend and blood oxygen saturation as the set of key vital signs indicators for the third stage, and extract the dynamic warning threshold of the set of key vital signs indicators for the third stage related to the expected duration of the operation. Query the nodes associated with the wound closure stage in the preset intraoperative nursing knowledge graph, obtain the mean arterial pressure and end-tidal carbon dioxide partial pressure as the set of key signs indicators for the fourth stage, and extract the dynamic warning threshold of the set of key signs indicators for the fourth stage related to the anesthesia recovery score.

[0009] Preferably, the step of mapping the discrete vital sign perception events to the set of key vital sign indicators to generate a phased vital sign status snapshot corresponding to the current phase type of the orthopedic surgery includes: Based on the current surgical stage code sent by the surgical anesthesia information system, determine the set of key vital signs currently in effect; Traverse all discrete vital sign perception events generated within the current time window and identify their vital sign type codes; The discrete vital sign perception events that match the vital sign type code with the indicator name defined in the currently effective set of key vital sign indicators are filtered out to form a subset of matching events. For multiple discrete vital sign perception events belonging to the same vital sign type in the subset of matching events, they are sorted by timestamp, and the current representative value of the vital sign type is calculated using a time-weighted average algorithm; Each indicator name in the currently effective set of key vital signs indicators is bound to its corresponding current representative value, and the latest timestamp of the discrete vital sign perception event is appended to it, together forming the phased vital sign status snapshot.

[0010] Preferably, the step of activating a preset nursing strategy reasoning engine and outputting composite nursing instructions based on the comparison results of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold includes: The current representative value of each vital sign data in the phased vital sign status snapshot is compared with the dynamic warning threshold of the corresponding item extracted from the preset intraoperative nursing knowledge graph. If the current representative value of one or more vital signs exceeds the dynamic warning threshold, a nursing decision trigger event is generated based on the type of vital sign exceeding the limit, the magnitude of the exceeding limit, and the current surgical stage code. The nursing decision triggering event is input into the preset nursing strategy reasoning engine, which performs reasoning based on the built-in rule tree. The node conditions of the rule tree are the logical AND operation between the combination of abnormal vital signs and the surgical stage, and the leaf nodes of the rule tree are associated with predefined nursing operation templates. After the preset nursing strategy reasoning engine matches the corresponding nursing operation template, it replaces the placeholders in the template with the current representative value, patient identifier and operating room number, and instantiates and generates a composite nursing instruction containing nursing resource identifier and execution parameters. The composite nursing instruction clearly specifies the identifier of the infusion pump to be called and the adjusted infusion rate.

[0011] Preferably, based on the composite nursing instructions, a virtual nursing task execution sequence is constructed within the AI-based orthopedic patient intraoperative nursing comprehensive management system, and the completion status of each task in the virtual nursing task execution sequence is tracked in real time, including: The composite nursing instruction is parsed to extract the nursing resource identifier and the execution parameters; Based on the nursing resource identifier, determine the physical equipment or nursing personnel roles required to execute the composite nursing instructions; Based on the execution parameters and the determined physical device or caregiver roles, a set of atomic care tasks with execution order dependencies are created; This group of atomic nursing tasks is sorted according to the execution order dependency to form the virtual nursing task execution sequence, and a unique status identifier is assigned to each atomic nursing task. Through the device communication interface or personnel terminal interface of the AI-based orthopedic patient intraoperative nursing integrated management system, control commands are sent to the corresponding physical devices or reminder information is sent to the corresponding personnel roles; The system continuously monitors for execution completion feedback signals from the physical device or confirmation operations from the personnel terminal, and updates the status identifier of the corresponding atomic nursing task in the virtual nursing task execution sequence based on the feedback signals or confirmation operations.

[0012] Preferably, the step of dynamically aggregating new discrete vital sign perception events from the heterogeneous sensing terminal based on the completion status of the virtual nursing task execution sequence, to iteratively update the phased vital sign status snapshot, includes: Monitor the status identifiers of all atomic nursing tasks in the virtual nursing task execution sequence; When the status indicator of a preset proportion of atomic nursing tasks changes to the completed status, a vital signs status snapshot update cycle is triggered. During the new vital sign status snapshot update cycle, the AI-based orthopedic patient intraoperative nursing integrated management system re-collects the discrete vital sign perception events newly generated since the last generation of the phased vital sign status snapshot. The same method as that used to generate the phased snapshot of the vital signs is employed, but new representative values ​​are calculated based on the newly collected discrete vital sign perception events and the current set of key vital sign indicators that may have changed as the surgical phase progresses. The corresponding value in the original phased vital sign state snapshot is replaced with the new current representative value, and the timestamp is updated, thereby completing the iterative update of the phased vital sign state snapshot.

[0013] Preferably, the method for calculating the RR interval, QRS complex amplitude and width for each heartbeat segment in the set of heartbeat segments to generate a standardized electrocardiogram event is as follows: For each heartbeat segment in the set of heartbeat segments, locate its R-wave peak point, and calculate the time interval between the R-wave peak point in the current heartbeat segment and the R-wave peak point in the previous heartbeat segment, which is taken as the RR interval. In each heartbeat segment, the Q wave start point and S wave end point are identified, and the time span from the Q wave start point to the S wave end point is calculated as the QRS complex width. In each heartbeat segment, the voltage difference from the baseline to the peak of the R wave is measured as the QRS complex amplitude. The calculated RR interval, QRS complex width, and QRS complex amplitude are encapsulated with the timestamp of the corresponding heartbeat segment and the source ECG lead information to generate the standardized ECG event.

[0014] Preferably, the method for obtaining the core body temperature change trend and blood oxygen saturation as the third-stage key vital sign indicator set, and extracting the dynamic warning threshold of the third-stage key vital sign indicator set related to the expected surgical duration, is as follows: From the pre-defined intraoperative nursing knowledge graph, retrieve physiological pathway nodes associated with key instrument operation stages; From the physiological pathway nodes, the core body temperature change trend indicators and blood oxygen saturation indicators are extracted and incorporated into the set of key vital signs indicators for the third stage. Obtain the estimated total duration set for the current surgery in the surgical scheduling system; Based on the estimated total duration, query the duration-threshold mapping relationship defined in the preset intraoperative nursing knowledge graph; Based on the duration-threshold mapping relationship, the allowable fluctuation range of core body temperature and the minimum safe limit of blood oxygen saturation corresponding to the expected total duration are obtained as the dynamic warning thresholds of the third stage key vital signs indicator set.

[0015] Preferably, the node conditions of the rule tree are a logical AND operation between the combination of abnormal vital signs and the surgical stage, and the leaf nodes of the rule tree are associated with the predefined nursing operation templates. The construction method is as follows: Define the root node of the rule tree as the nursing decision trigger event; In each non-leaf node of the rule tree, a judgment condition is set. The judgment condition is the result of the logical AND operation between the out-of-limit state of a specific combination of vital sign data and the code of a specific surgical stage. Starting from the root node, based on the excess status of vital signs data in the phased vital sign status snapshot and the current surgical stage code, the branch path that meets the conditions is traversed downwards layer by layer. When traversing to a leaf node, the leaf node is bound to a pre-configured nursing operation template for a specific combination of abnormal signs and a specific surgical stage scenario represented by the branch path. The nursing operation template includes a framework of nursing operation types to be performed, operation object identifiers, and parameter setting instructions.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Based on the different stages of the surgery, a set of key vital signs and their dynamic warning thresholds are extracted from a pre-defined intraoperative nursing knowledge graph. This allows the system to identify the specific step in the current surgical process and automatically apply the optimal monitoring strategy for that step. The evaluation criteria for vital signs are no longer fixed but dynamically adjusted according to the risk changes during the surgical process. This improves the accuracy and clinical relevance of warnings, reduces disruptive alarms in non-critical stages, and enables earlier and more accurate detection of abnormal physiological trends in high-risk stages, transforming monitoring from uniform surveillance to proactive perception synchronized with the surgical rhythm.

[0017] Based on the comparison between periodic vital sign status snapshots and dynamic thresholds, the nursing strategy inference engine is activated to output composite nursing instructions. Subsequently, a virtual nursing task execution sequence is constructed within the system, tracking the completion status of each sub-task in real time. This transforms a single alarm signal into a structured action plan containing multiple ordered steps. The system not only identifies "what the problem is" but also infers "how to systematically address it." By constructing and tracking the virtual task sequence, the nursing team's various response actions are incorporated into a manageable framework. This achieves structured and collaborative nursing interventions, avoiding omissions or conflicts in treatment measures. The system can iteratively update the status snapshot based on task completion status and new vital sign perception events, thus forming a closed loop of decision-making, execution, feedback, and re-decision-making. This allows the effectiveness of nursing actions to be evaluated immediately, and subsequent strategies to be dynamically adjusted based on the patient's real-time response, achieving a leap from one-off responses to continuous, adaptive management. Attached Figure Description

[0018] Figure 1 This is a sequence diagram of the artificial intelligence-based intraoperative nursing management system for orthopedic patients described in this invention. Figure 2 A flowchart generated for analyzing vital sign perception events; Figure 3 A flowchart for generating snapshots of phased vital signs; Figure 4 Stacked bar chart showing the frequency of abnormal physical signs at different stages of orthopedic surgery; Figure 5 Radar graphs showing the stability of vital signs at different stages of orthopedic surgery. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0021] See Figure 1This invention provides an artificial intelligence-based comprehensive management system for intraoperative nursing care of orthopedic patients. This system achieves closed-loop management of intelligent perception, analysis, decision-making, and nursing intervention of patient vital signs during surgery. The system's event parsing module first acquires raw monitoring signals of patient vital signs from various heterogeneous sensing terminals in the operating room, such as electrocardiogram monitors, non-invasive blood pressure monitors, and temperature sensors. This module parses and standardizes these raw monitoring signals from different sources and in different formats, transforming them into data units with a unified data structure, i.e., discrete vital sign perception events. The rule extraction module, based on the current surgical stage type obtained from the surgical anesthesia information system (e.g., anesthesia induction, surgical incision), extracts a set of key vital sign indicators associated with the stage type from a preset intraoperative nursing knowledge graph, and obtains the dynamic warning threshold for each key vital sign indicator in this set. The state snapshot module maps the real-time generated discrete vital sign perception events to the currently effective set of key vital sign indicators, and generates a phased vital sign state snapshot reflecting the patient's current physiological state through calculation. The strategy decision-making module compares the vital sign data in the status snapshot with their corresponding dynamic warning thresholds. When a threshold is exceeded, a preset nursing strategy reasoning engine is activated to perform logical reasoning and output a composite nursing instruction containing specific operational instructions. The sequence control module parses the composite nursing instruction, breaks it down into ordered atomic tasks, constructs a virtual nursing task execution sequence within the system, and tracks the execution and completion status of each task in the sequence in real time. The iterative update module triggers an update cycle based on the completion status of the virtual nursing task execution sequence, re-aggregates the latest discrete vital sign perception events, and uses the updated data and potentially changing key vital sign indicators to iteratively update the periodic vital sign status snapshot, thereby initiating a new round of monitoring-decision-execution cycle.

[0022] In one embodiment of the present invention, see [reference] Figure 2 The event analysis module acquires raw monitoring signals of the patient's vital signs from multiple heterogeneous sensing terminals within the surgical environment. These raw monitoring signals include analog voltage waveform sequences from the electrocardiogram monitor, inflation / deflation pressure data streams from the non-invasive blood pressure monitor, and impedance change values ​​from the body temperature sensor. The event analysis module performs heartbeat segmentation based on periodic characteristics on the analog voltage waveform sequence, identifying the start and end points of each heartbeat cycle, thus dividing the continuous analog voltage waveform sequence into independent sets of heartbeat segments. It performs blood pressure cycle identification based on pressure inflection points on the inflation / deflation pressure data stream, dividing independent blood pressure measurement cycles by detecting inflection points of pressure rises and falls in the inflation / deflation pressure data stream. Finally, it performs mean filtering based on a time window on the impedance change values, averaging the impedance change values ​​using a sliding window of fixed time length to obtain filtered body temperature time-series data.

[0023] For each heartbeat segment in the heartbeat segment set, the event parsing module calculates the RR interval, QRS complex amplitude, and width to generate a standardized ECG event. The RR interval is calculated by locating the R-wave peak in the current heartbeat segment and calculating the time interval between the R-wave peak in the current segment and the R-wave peak in the previous segment. The QRS complex width is calculated by identifying the Q-wave start and S-wave end points in the heartbeat segment and calculating the time span from the Q-wave start to the S-wave end. The QRS complex amplitude is calculated by measuring the voltage difference from the baseline to the R-wave peak in the heartbeat segment. In practice, the QRS complex width can be calculated using the following formula:

[0024] in: Represents the QRS group width. The timestamp representing the end of the S-wave. The timestamp represents the start of the Q wave. The event parsing module encapsulates the calculated RR interval, QRS complex width, and QRS complex amplitude with the timestamp of the corresponding heartbeat segment and the source ECG lead information to generate a standardized ECG event. Optionally, the source ECG lead information includes lead location identifiers and sampling frequency.

[0025] For each independent blood pressure measurement cycle, the event analysis module extracts the peak systolic blood pressure, the trough diastolic blood pressure, and the mean arterial pressure (MAP) to generate a standardized blood pressure event. The method for extracting the peak systolic blood pressure is to find the pressure value corresponding to the highest point of the pressure curve in the blood pressure measurement cycle; the method for extracting the trough diastolic blood pressure is to find the pressure value corresponding to the lowest point of the pressure curve in the blood pressure measurement cycle; the method for calculating the MAP is a weighted average based on the peak systolic blood pressure and the trough diastolic blood pressure. In practice, the MAP can be calculated using the formula:

[0026] in: Represents mean arterial pressure. Represents peak systolic blood pressure. This represents the trough value of diastolic blood pressure. The event parsing module encapsulates the peak systolic blood pressure, trough diastolic blood pressure, and mean arterial pressure into a standardized blood pressure event.

[0027] For the filtered body temperature time-series data, the event parsing module samples at fixed time intervals to generate standardized body temperature events. The fixed time interval is a preset time period. The event parsing module extracts temperature values ​​from the filtered body temperature time-series data at fixed time intervals and binds the timestamp of each sampling point with the temperature value to form a standardized body temperature event. The fixed time interval can be configured according to monitoring needs, for example, set to 1 second or 5 seconds. The event parsing module encapsulates standardized ECG events, standardized blood pressure events, and standardized body temperature events into a discrete vital sign sensing event containing a timestamp, vital sign type code, numerical value, and source identifier. In specific implementation, the vital sign type code is a predefined enumerated value used to distinguish vital sign types such as heart rate, blood pressure, and body temperature; the source identifier is used to identify the sensing terminal device from which the data originates.

[0028] In one embodiment of the present invention, see [reference] Figure 3 The rule extraction module extracts key vital signs (KV) sets associated with different stages of orthopedic surgery from a pre-defined intraoperative nursing knowledge graph, along with dynamic warning thresholds for each KV indicator. The module also receives the current surgical stage code from the surgical anesthesia information system. This code includes the anesthesia induction stage, surgical incision stage, key instrument operation stage, and wound closure stage. In practice, the surgical anesthesia information system pushes the current surgical stage code to the rule extraction module in real-time via a medical data interface. The surgical stage code can be identified using predefined enumeration strings such as "STAGE_01" and "STAGE_02".

[0029] The rule extraction module queries a pre-defined intraoperative nursing knowledge graph for nodes associated with the anesthesia induction phase, obtaining heart rate variability and blood pressure drop rate as the first-stage key vital signs indicator set, and extracting dynamic warning thresholds for the first-stage key vital signs indicator set related to age and baseline blood pressure. In some embodiments, the pre-defined intraoperative nursing knowledge graph is stored in a graph database. The rule extraction module executes a graph query language, starting from the node identified as "anesthesia induction," traversing the "monitoring focus" edges to obtain associated heart rate variability and blood pressure drop rate entity nodes. The rule extraction module further reads the patient's age and preoperative baseline blood pressure values ​​from the patient's electronic medical record, using these as input parameters to query predefined logical rules in the pre-defined intraoperative nursing knowledge graph, calculating specific warning thresholds for heart rate variability and blood pressure drop rate. These specific warning thresholds constitute the dynamic warning thresholds for the first-stage key vital signs indicator set.

[0030] The rule extraction module queries the pre-defined intraoperative nursing knowledge graph for nodes associated with the surgical incision stage, obtains the instantaneous rise in heart rate and blood pressure as the set of key vital signs indicators for the second stage, and extracts the dynamic warning thresholds of the set of key vital signs indicators for the second stage related to the patient's pain threshold. It can be understood that the "pain sensitivity" attribute associated with the "surgical incision" node in the pre-defined intraoperative nursing knowledge graph affects the dynamic warning thresholds for the instantaneous rise in heart rate and blood pressure. The rule extraction module matches the threshold ranges corresponding to different levels in the pre-defined intraoperative nursing knowledge graph based on the pain threshold classification obtained from the patient's assessment records.

[0031] The rule extraction module queries a pre-defined intraoperative nursing knowledge graph for nodes associated with key instrument operation stages, obtaining core body temperature change trends and blood oxygen saturation as the set of key vital signs indicators for the third stage. In specific implementations, it retrieves physiological path nodes associated with key instrument operation stages from the pre-defined intraoperative nursing knowledge graph, extracts core body temperature change trend indicators and blood oxygen saturation indicators from these physiological path nodes, and incorporates them into the set of key vital signs indicators for the third stage. The rule extraction module obtains the estimated total duration set for the current surgery in the surgical scheduling system. Based on the estimated total duration, the rule extraction module queries the duration-threshold mapping relationship defined in the pre-defined intraoperative nursing knowledge graph. According to the duration-threshold mapping relationship, it obtains the allowable fluctuation range of core body temperature and the minimum safe limit of blood oxygen saturation corresponding to the estimated total duration, serving as the dynamic warning thresholds for the set of key vital signs indicators for the third stage. In some embodiments, the calculation of the allowable fluctuation range of core body temperature is positively correlated with the estimated total duration. Optionally, the duration-threshold mapping relationship is stored in a pre-defined intraoperative nursing knowledge graph in the form of a lookup table. The rule extraction module indexes this table by the expected total duration and directly reads the corresponding core allowable upper limit of body temperature fluctuation. and lower limit value and minimum safe limit for blood oxygen saturation .in This represents the maximum permissible positive fluctuation in core body temperature. This represents the maximum permissible negative fluctuation in core body temperature. This represents the minimum permissible value for blood oxygen saturation. The rule extraction module queries the nodes associated with the wound closure stage in the preset intraoperative nursing knowledge graph, obtains the mean arterial pressure and end-tidal carbon dioxide partial pressure as the set of key signs for the fourth stage, and extracts the dynamic warning threshold of the set of key signs for the fourth stage related to the anesthesia recovery score.

[0032] In one embodiment of the present invention, the state snapshot module maps discrete vital sign perception events to a set of key vital sign indicators to generate a phased vital sign state snapshot corresponding to the current phase type of orthopedic surgery. The state snapshot module determines the currently effective set of key vital sign indicators based on the current surgical phase code sent by the surgical anesthesia information system. The current surgical phase code is pushed in real time by the surgical anesthesia information system when the surgical process changes, for example, switching from the "anesthesia induction" phase code to the "surgical incision" phase code. After receiving the new current surgical phase code, the state snapshot module requests the rule extraction module to obtain the set of key vital sign indicators corresponding to the current surgical phase code. This set of key vital sign indicators is then marked as the currently effective set of key vital sign indicators.

[0033] The state snapshot module iterates through all discrete vital sign sensing events generated within the current time window and identifies their vital sign type codes. In specific implementations, the current time window is a configurable sliding time interval, such as the most recent 30 seconds or the most recent 1 minute. The state snapshot module retrieves all discrete vital sign sensing events whose timestamps fall within the current time window from the output buffer of the event parsing module. The state snapshot module parses the predefined "vital sign type code" field in each discrete vital sign sensing event. This field identifies that the event represents a specific vital sign type such as heart rate, blood pressure, body temperature, or blood oxygen saturation. The state snapshot module filters out discrete vital sign sensing events whose vital sign type codes match the indicator names defined in the currently effective set of key vital sign indicators, forming a subset of matching events. In some embodiments, the currently effective set of key vital sign indicators is a list, where each item contains a vital sign indicator name, such as "heart rate variability" or "core body temperature change trend". The state snapshot module compares the vital sign type code of discrete vital sign perception events with the name of each vital sign indicator in the list. Discrete vital sign perception events whose vital sign type code matches any vital sign indicator name are added to the matching event subset, while events whose vital sign type code is not in the currently effective set of key vital sign indicators are temporarily ignored.

[0034] For multiple discrete vital sign perception events belonging to the same vital sign type within the matched event subset, the state snapshot module sorts them by timestamp and calculates the current representative value of the vital sign type using a time-weighted average algorithm. In specific implementation, for a particular vital sign type, the state snapshot module finds all events of that type from the matched event subset and sorts them in ascending order according to the chronological order of their timestamp fields. The time-weighted average algorithm assigns higher weights to events closer to the current moment, resulting in a higher current representative value. The calculation can be performed using the formula:

[0035] in: This represents the current representative value obtained through calculation. This represents the total number of events of that trait type within the subset of matched events. Representing the Vital sign values ​​in each event. Representing the The timestamp of each event Represents the current system time. It is a decay coefficient between 0 and 1, used to control the rate at which the time weight decays.

[0036] Each indicator name in the currently effective set of key vital signs indicators is bound to its corresponding current representative value, and the latest timestamp of the discrete vital sign perception event is appended to form a phased vital sign status snapshot. In some embodiments, the phased vital sign status snapshot is a structured data object containing a list. Each entry in the list consists of three fields: field 1 is the indicator name extracted from the set of key vital signs indicators; field 2 is the current representative value corresponding to the indicator calculated by the time-weighted average algorithm described above; and field 3 is the latest timestamp among all relevant discrete vital sign perception events used to calculate the current representative value. Optionally, if a key vital sign indicator does not have a corresponding discrete vital sign perception event within the current time window, its current representative value field is marked as "data missing," and the latest timestamp field records the timestamp of the last valid update or is empty.

[0037] In one embodiment of the present invention, the strategy decision module activates a preset nursing strategy reasoning engine and outputs composite nursing instructions based on the comparison results of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold. The strategy decision module compares the current representative value of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold of the corresponding item extracted from the preset intraoperative nursing knowledge graph. In specific implementation, the phased vital sign status snapshot is input into the strategy decision module in the form of a data structure. The strategy decision module reads the name of each vital sign indicator and its corresponding current representative value in the data structure, and at the same time obtains the dynamic warning threshold of each vital sign indicator in the current stage from the output of the rule extraction module. The dynamic warning threshold may be a single value range, or it may be an interval defined by an upper limit and a lower limit. The strategy decision module executes the numerical comparison logic.

[0038] If the current representative value of one or more vital signs exceeds the dynamic warning threshold, a nursing decision trigger event is generated based on the type of vital sign exceeding the limit, the magnitude of the exceedance, and the current surgical stage code. The nursing decision trigger event is a structured data object containing the following fields: trigger timestamp, a list of vital sign names that triggered the trigger, the current representative value and dynamic warning threshold for each vital sign, the direction and magnitude of the exceedance for each vital sign, and the current surgical stage code. The magnitude of the exceedance can be calculated based on the ratio of the difference between the current representative value and the dynamic warning threshold boundary value; for example, for cases exceeding the upper limit, the magnitude of the exceedance... formula:

[0039] in: Represents the percentage of the exceedance. The current representative value of the vital signs data, This represents the upper limit of the dynamic warning threshold for this vital sign indicator. It's understandable that for values ​​below the lower limit, the formula needs to be adjusted accordingly to calculate the difference. Refer to Table 1, which shows a comparison of core data that may be included in a nursing decision-triggered event triggered during the surgical incision stage.

[0040] Table 1: Comparison of Vital Signs Data and Dynamic Alert Thresholds Name of vital signs Current representative values Dynamic warning threshold Over-limit direction Calculation value of over-limit amplitude instantaneous increase in heart rate +28 times / minute ≤20 beats / minute Exceeding the limit +40% Instantaneous rise in blood pressure +35mmHg ≤+25mmHg Exceeding the limit +40% Core body temperature change trend -0.2°C [-0.5°C, +0.5°C] Not exceeded 0% The strategy decision-making module inputs nursing decision trigger events into a pre-defined nursing strategy reasoning engine. This engine reasones based on a built-in rule tree, with the root node of the rule tree defined as the nursing decision trigger event. At each non-leaf node of the rule tree, a judgment condition is set: the result of a logical AND operation between the out-of-limit state of a specific combination of vital sign data and the code for a specific surgical stage. Starting from the root node, the pre-defined nursing strategy reasoning engine traverses downwards layer by layer, judging the out-of-limit state of vital sign data in the nursing decision trigger event and the current surgical stage code, selecting the branch path that meets the condition. The node conditions of the rule tree are the logical AND operation between the combination of vital sign out-of-limits and the surgical stage. The leaf nodes of the rule tree are associated with predefined nursing operation templates. When traversing to a leaf node, the leaf node is bound to a pre-configured nursing operation template for the specific combination of vital sign out-of-limits and the specific surgical stage scenario represented by the branch path. The nursing operation template contains a framework of the nursing operation type to be performed, the operation object identifier, and parameter setting instructions. After the pre-defined nursing strategy reasoning engine matches the corresponding nursing operation template, it replaces the placeholders in the template with the current representative values, patient identifier, and operating room number from the nursing decision trigger event. It then instantiates and generates a composite nursing instruction containing nursing resource identifiers and execution parameters. This composite nursing instruction explicitly specifies the identifier of the infusion pump to be invoked and the adjusted infusion rate. In essence, the composite nursing instruction is a serializable instruction object containing complete operational semantics.

[0041] See Figure 4 This is a stacked bar chart showing the frequency of abnormal vital signs at different stages of orthopedic surgery. This chart is part of a medical intraoperative nursing data visualization tool, used to visually display the risk distribution of vital signs at different surgical stages, helping medical staff quickly identify high-risk stages (such as surgical skin incision) and prepare for nursing interventions in advance. It intuitively presents the types of high-risk vital signs at each surgical stage, helping medical staff to focus on monitoring relevant indicators at the corresponding stage and prepare intervention plans in advance; by identifying the differences in trigger frequency at each stage, it identifies "high-load links" in intraoperative nursing (such as surgical skin incision), allowing for targeted optimization of the nursing process at that stage; it can also serve as one of the quantitative indicators of intraoperative nursing quality, comparing the differences in trigger frequency among different patients and different medical teams, and assisting in evaluating the effectiveness of nursing measures.

[0042] In one embodiment of the present invention, the iterative update module dynamically aggregates new discrete vital sign perception events from heterogeneous sensing terminals based on the completion status of the virtual nursing task execution sequence to iteratively update the phased vital sign status snapshot. This module monitors the status identifiers of all atomic nursing tasks in the virtual nursing task execution sequence. When a preset proportion of the atomic nursing task status identifiers changes to the completed state, a vital sign status snapshot update cycle is triggered. Within the new vital sign status snapshot update cycle, the AI-based orthopedic patient intraoperative nursing comprehensive management system re-collects the newly generated discrete vital sign perception events since the last generation of the phased vital sign status snapshot. Using the same method as generating the phased vital sign status snapshot, but based on the newly collected discrete vital sign perception events and the current set of key vital sign indicators that may have changed due to the progress of the surgical stage, a new current representative value is calculated and generated. The new current representative value replaces the corresponding value in the original phased vital sign status snapshot, and the timestamp is updated, thereby completing the iterative update of the phased vital sign status snapshot.

[0043] In practical implementation, the sequence control module constructs a virtual nursing task execution sequence within the AI-based intraoperative nursing management system for orthopedic patients, based on composite nursing instructions. It tracks the completion status of each task in this virtual sequence in real time. The sequence control module parses the composite nursing instructions, extracting nursing resource identifiers and execution parameters. In practice, a composite nursing instruction is a data object containing structured fields. The sequence control module obtains the nursing resource identifier by parsing the "targetDevice" or "targetRole" field and the execution parameters by parsing the "actionParameters" field. The nursing resource identifier can be a specific device network address or personnel role code, while the execution parameters include specific control command values ​​or prompt information.

[0044] The sequence control module determines the physical devices or nursing personnel roles required to execute complex nursing instructions based on nursing resource identifiers. In some embodiments, the sequence control module maintains a device and role registry that maps nursing resource identifiers to specific device communication interfaces or personnel-held terminal applications. When a nursing resource identifier points to an infusion pump, the sequence control module queries the registry for the infusion pump's IP address and communication protocol. When a nursing resource identifier points to the "anesthesia nurse" role, the sequence control module queries the registry for the mobile terminal session identifier logged in by the personnel currently assuming this role in the operating room.

[0045] Based on execution parameters and defined physical devices or nursing personnel roles, the sequence control module creates a set of atomic nursing tasks with execution order dependencies. It can be understood that a complex nursing instruction may need to be broken down into multiple steps, where each subsequent task can only begin after the previous one has been successfully completed. According to predefined task decomposition rules, the sequence control module transforms the execution parameters into a series of ordered atomic operation descriptions and labels each atomic nursing task with its execution object as a specific physical device or personnel role. This set of atomic nursing tasks is then sorted according to their execution order dependencies, forming a virtual nursing task execution sequence, and each atomic nursing task is assigned a unique status identifier. The virtual nursing task execution sequence exists in memory as a linked list or array data structure. Each node represents an atomic nursing task, and the node contains a task description, execution object, execution parameters, order constraints, and a status identifier field. The initial value of the status identifier is set to "pending execution".

[0046] Through the device communication interface or personnel terminal interface of the AI-based orthopedic patient intraoperative nursing integrated management system, the sequence control module sends control commands to the corresponding physical devices or reminder information to the corresponding personnel roles. The device communication interface conforms to medical device communication standards. The sequence control module encodes the control commands contained in the atomic nursing tasks into device-recognizable instruction messages and sends them over the network. The personnel terminal interface pushes text, voice, or visual reminder information to the terminal devices used by nursing staff via message queues or instant messaging protocols. The system continuously monitors execution completion feedback signals from physical devices or confirmation operations from personnel terminals, and updates the status identifier of the corresponding atomic nursing task in the virtual nursing task execution sequence based on these signals or confirmation operations. In specific implementation, after executing a control command, the physical device sends back a confirmation message containing the execution result code through the device communication interface. After receiving the reminder information, the operator clicks the confirmation button to generate a confirmation event. The sequence control module monitors these feedback signals and confirmation events and matches them with the atomic nursing tasks in the virtual nursing task execution sequence. If a match is successful, the status identifier of the corresponding atomic nursing task is updated to "completed" or "execution failed."

[0047] The iterative update module dynamically aggregates new discrete vital sign sensing events from heterogeneous sensing terminals based on the completion status of the virtual nursing task execution sequence, iteratively updating the phased vital sign status snapshot. The iterative update module monitors the status identifiers of all atomic nursing tasks in the virtual nursing task execution sequence. In specific implementations, the iterative update module subscribes to event notifications from the sequence control module regarding changes in the status of the virtual nursing task execution sequence. When a status change event occurs, the iterative update module obtains the latest list of status identifiers for all atomic nursing tasks in the current virtual nursing task execution sequence. When a preset percentage of atomic nursing tasks changes their status identifier to "completed," a vital sign status snapshot update cycle is triggered. The preset percentage is a configurable threshold, such as 80% or 100%. The iterative update module calculates the ratio of the number of atomic nursing tasks with the status identifier "completed" to the total number of atomic nursing tasks. When the ratio reaches or exceeds the preset percentage, the triggering condition is met. In some embodiments, the triggering condition can be expressed by the formula... It means that, among them This represents the completion rate of a virtual nursing task execution sequence, and its value is the number of atomic nursing tasks that have been completed. Total number of atomic nursing tasks The merchant, that is , This represents a preset ratio threshold.

[0048] Within the new vital sign status snapshot update cycle, the AI-based orthopedic patient intraoperative nursing integrated management system re-collects newly generated discrete vital sign perception events since the last generation of the staged vital sign status snapshot. This re-collection process involves querying the output cache or message queue of the event parsing module for all discrete vital sign perception events with timestamps later than the latest timestamp recorded in the last staged vital sign status snapshot. These events constitute a new dataset for updating. Using the same method as generating the staged vital sign status snapshot, but based on the newly collected discrete vital sign perception events and the current set of key vital sign indicators that may have changed due to the progress of the surgical stage, a new set of current representative values ​​is calculated. The iterative update module first obtains the latest set of key vital sign indicators from the rule extraction module, and then calculates a new set of current representative values ​​using the newly collected discrete vital sign perception events, following the mapping and time-weighted averaging algorithm implemented by the status snapshot module. The corresponding values ​​in the original phased vital sign state snapshot are replaced with new current representative values, and the timestamps are updated to complete the iterative update of the phased vital sign state snapshot. The iterative update module generates a new phased vital sign state snapshot data object. The list of indicator names in this object is consistent with the set of key vital sign indicators. The numerical fields of each indicator are updated to the new current representative values ​​obtained by calculation, and the latest timestamp field is updated to the latest timestamp of the newly collected discrete vital sign perception events used in this calculation.

[0049] See Figure 5 This is a radar chart of vital sign stability at various stages of orthopedic surgery, used for multi-dimensional comparison of physiological indicators at different surgical stages. It visually presents the strengths and weaknesses of vital signs at different stages, helping medical staff to adjust monitoring priorities accordingly; quickly pinpoints "weaknesses in certain stages," allowing for early deployment of nursing interventions; and serves as a comprehensive quantitative tool for intraoperative nursing effectiveness, assisting in optimizing the phased strategies of the "Orthopedic Intraoperative Nursing Management System." Through multi-dimensional comparison, it quickly identifies "vital sign weaknesses" at each stage, allowing medical staff to focus on strengthening monitoring and intervention in that dimension at the corresponding stage. By combining the patient's actual condition with the differences between the stability of vital signs at each stage and the radar chart baseline data, more precise intraoperative nursing plans can be developed. This provides a quantitative basis for the phased rules of the "Orthopedic Intraoperative Nursing Management System."

[0050] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An artificial intelligence-based comprehensive management system for intraoperative nursing care of orthopedic patients, characterized in that: The system includes: The event parsing module acquires raw monitoring signals about the patient's vital signs from multiple heterogeneous sensing terminals within the surgical environment, and parses the raw monitoring signals into discrete vital sign sensing events with a unified data structure. The rule extraction module extracts a set of key signs associated with different stages of orthopedic surgery from a preset intraoperative nursing knowledge graph, as well as the dynamic warning threshold of each key sign in the set of key signs. The status snapshot module maps the discrete vital sign perception events to the set of key vital sign indicators to generate a staged vital sign status snapshot corresponding to the current stage type of the orthopedic surgery. The strategy decision-making module, based on the comparison results of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold, activates the preset nursing strategy reasoning engine and outputs composite nursing instructions. The sequence control module, based on the composite nursing instructions, constructs a virtual nursing task execution sequence within the AI-based orthopedic patient intraoperative nursing comprehensive management system, and tracks the completion status of each task in the virtual nursing task execution sequence in real time. The iterative update module dynamically aggregates new discrete vital sign perception events from the heterogeneous sensing terminal based on the completion status of the virtual nursing task execution sequence, so as to iteratively update the phased vital sign status snapshot.

2. The comprehensive management system for intraoperative nursing care of orthopedic patients based on artificial intelligence according to claim 1, characterized in that, The process of parsing the original monitoring signal into discrete vital sign perception events with a unified data structure includes: The system receives an analog voltage waveform sequence from an electrocardiogram monitor, an inflation / deflation pressure data stream from a non-invasive blood pressure monitor, and an impedance change value from a body temperature sensor. The analog voltage waveform sequence, the inflation / deflation pressure data stream, and the impedance change value together constitute the original monitoring signal. The analog voltage waveform sequence is subjected to heartbeat segmentation based on periodic features, the inflation and deflation pressure data stream is subjected to blood pressure cycle identification based on pressure inflection points, and the impedance change value is subjected to mean filtering based on time windows to obtain a set of heartbeat segments, independent blood pressure measurement cycles, and filtered body temperature time series data, respectively. For each heartbeat segment in the aforementioned heartbeat segment set, the RR interval, QRS complex amplitude and width are calculated to generate a standardized electrocardiogram event; For each independent blood pressure measurement cycle, the peak systolic blood pressure, the trough diastolic blood pressure, and the mean arterial pressure are extracted to generate standardized blood pressure events. The filtered body temperature time-series data is sampled at fixed time intervals to generate standardized body temperature events; The standardized electrocardiogram event, the standardized blood pressure event, and the standardized body temperature event are uniformly encapsulated into a discrete vital sign perception event containing a timestamp, vital sign type code, numerical value, and source identifier.

3. The comprehensive management system for intraoperative nursing care of orthopedic patients based on artificial intelligence according to claim 2, characterized in that, The process involves extracting a set of key vital signs associated with different stages of orthopedic surgery from a pre-defined intraoperative nursing knowledge graph, along with a dynamic warning threshold for each key vital sign in the set. This includes: Receive the current surgical stage code sent by the surgical anesthesia information system. The surgical stage code includes the anesthesia induction stage, surgical incision stage, key instrument operation stage, and wound closure stage. Query the nodes associated with the anesthesia induction phase in the preset intraoperative nursing knowledge graph, obtain heart rate variability and blood pressure drop rate as the first phase key sign indicator set, and extract the dynamic warning threshold of the first phase key sign indicator set related to age and baseline blood pressure. Query the nodes associated with the surgical incision stage in the preset intraoperative nursing knowledge graph, obtain the instantaneous increase in heart rate and blood pressure as the second stage key sign indicator set, and extract the dynamic warning threshold of the second stage key sign indicator set related to the patient's pain threshold. Query the nodes associated with the key instrument operation stage in the preset intraoperative nursing knowledge graph, obtain the core body temperature change trend and blood oxygen saturation as the set of key vital signs indicators for the third stage, and extract the dynamic warning threshold of the set of key vital signs indicators for the third stage related to the expected duration of the operation. Query the nodes associated with the wound closure stage in the preset intraoperative nursing knowledge graph, obtain the mean arterial pressure and end-tidal carbon dioxide partial pressure as the set of key signs indicators for the fourth stage, and extract the dynamic warning threshold of the set of key signs indicators for the fourth stage related to the anesthesia recovery score.

4. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 3, characterized in that, The step of mapping the discrete vital sign perception events to the set of key vital sign indicators to generate a phased vital sign status snapshot corresponding to the current phase type of the orthopedic surgery includes: Based on the current surgical stage code sent by the surgical anesthesia information system, determine the set of key vital signs currently in effect; Traverse all discrete vital sign perception events generated within the current time window and identify their vital sign type codes; The discrete vital sign perception events that match the vital sign type code with the indicator name defined in the currently effective set of key vital sign indicators are filtered out to form a subset of matching events. For multiple discrete vital sign perception events belonging to the same vital sign type in the subset of matching events, they are sorted by timestamp, and the current representative value of the vital sign type is calculated using a time-weighted average algorithm; Each indicator name in the currently effective set of key vital signs indicators is bound to its corresponding current representative value, and the latest timestamp of the discrete vital sign perception event is appended to it, together forming the phased vital sign status snapshot.

5. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 4, characterized in that, Based on the comparison results of each vital sign data in the phased vital sign status snapshot with the dynamic warning threshold, a preset nursing strategy reasoning engine is activated to output a composite nursing instruction; including: The current representative value of each vital sign data in the phased vital sign status snapshot is compared with the dynamic warning threshold of the corresponding item extracted from the preset intraoperative nursing knowledge graph. If the current representative value of one or more vital signs exceeds the dynamic warning threshold, a nursing decision trigger event is generated based on the type of vital sign exceeding the limit, the magnitude of the exceeding limit, and the current surgical stage code. The nursing decision triggering event is input into the preset nursing strategy reasoning engine, which performs reasoning based on the built-in rule tree. The node conditions of the rule tree are the logical AND operation between the combination of abnormal vital signs and the surgical stage, and the leaf nodes of the rule tree are associated with predefined nursing operation templates. After the preset nursing strategy reasoning engine matches the corresponding nursing operation template, it replaces the placeholders in the template with the current representative value, patient identifier and operating room number, and instantiates and generates a composite nursing instruction containing nursing resource identifier and execution parameters. The composite nursing instruction clearly specifies the identifier of the infusion pump to be called and the adjusted infusion rate.

6. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 5, characterized in that, Based on the composite nursing instructions, a virtual nursing task execution sequence is constructed within the AI-based orthopedic patient intraoperative nursing comprehensive management system, and the completion status of each task in the virtual nursing task execution sequence is tracked in real time, including: The composite nursing instruction is parsed to extract the nursing resource identifier and the execution parameters; Based on the nursing resource identifier, determine the physical equipment or nursing personnel roles required to execute the composite nursing instructions; Based on the execution parameters and the determined physical device or caregiver roles, a set of atomic care tasks with execution order dependencies are created; This group of atomic nursing tasks is sorted according to the execution order dependency to form the virtual nursing task execution sequence, and a unique status identifier is assigned to each atomic nursing task. Through the device communication interface or personnel terminal interface of the AI-based orthopedic patient intraoperative nursing integrated management system, control commands are sent to the corresponding physical devices or reminder information is sent to the corresponding personnel roles; The system continuously monitors for execution completion feedback signals from the physical device or confirmation operations from the personnel terminal, and updates the status identifier of the corresponding atomic nursing task in the virtual nursing task execution sequence based on the feedback signals or confirmation operations.

7. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 6, characterized in that, The step of dynamically aggregating new discrete vital sign perception events from the heterogeneous sensing terminal based on the completion status of the virtual nursing task execution sequence, in order to iteratively update the phased vital sign status snapshot, includes: Monitor the status identifiers of all atomic nursing tasks in the virtual nursing task execution sequence; When the status indicator of a preset proportion of atomic nursing tasks changes to the completed status, a vital signs status snapshot update cycle is triggered. During the new vital sign status snapshot update cycle, the AI-based orthopedic patient intraoperative nursing integrated management system re-collects the discrete vital sign perception events newly generated since the last generation of the phased vital sign status snapshot. The same method as that used to generate the phased snapshot of the vital signs is employed, but new representative values ​​are calculated based on the newly collected discrete vital sign perception events and the current set of key vital sign indicators that may have changed as the surgical phase progresses. The corresponding value in the original phased vital sign state snapshot is replaced with the new current representative value, and the timestamp is updated, thereby completing the iterative update of the phased vital sign state snapshot.

8. The comprehensive management system for intraoperative nursing care of orthopedic patients based on artificial intelligence according to claim 2, characterized in that, The method for calculating the RR interval, QRS complex amplitude and width for each heartbeat segment in the set of heartbeat segments, and generating a standardized electrocardiogram event is as follows: For each heartbeat segment in the set of heartbeat segments, locate its R-wave peak point, and calculate the time interval between the R-wave peak point in the current heartbeat segment and the R-wave peak point in the previous heartbeat segment, which is taken as the RR interval. In each heartbeat segment, the Q wave start point and S wave end point are identified, and the time span from the Q wave start point to the S wave end point is calculated as the QRS complex width. In each heartbeat segment, the voltage difference from the baseline to the peak of the R wave is measured as the QRS complex amplitude. The calculated RR interval, QRS complex width, and QRS complex amplitude are encapsulated with the timestamp of the corresponding heartbeat segment and the source ECG lead information to generate the standardized ECG event.

9. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 3, characterized in that, The method for obtaining the core body temperature change trend and blood oxygen saturation as the third-stage key vital signs indicator set, and extracting the dynamic warning threshold of the third-stage key vital signs indicator set related to the expected operation time, is as follows: From the pre-defined intraoperative nursing knowledge graph, retrieve physiological pathway nodes associated with key instrument operation stages; From the physiological pathway nodes, the core body temperature change trend indicators and blood oxygen saturation indicators are extracted and incorporated into the set of key vital signs indicators for the third stage. Obtain the estimated total duration set for the current surgery in the surgical scheduling system; Based on the estimated total duration, query the duration-threshold mapping relationship defined in the preset intraoperative nursing knowledge graph; Based on the duration-threshold mapping relationship, the allowable fluctuation range of core body temperature and the minimum safe limit of blood oxygen saturation corresponding to the expected total duration are obtained as the dynamic warning thresholds of the third stage key vital signs indicator set.

10. The artificial intelligence-based intraoperative nursing management system for orthopedic patients according to claim 5, characterized in that, The node conditions of the rule tree are a logical AND operation between the combination of abnormal vital signs and the surgical stage. The leaf nodes of the rule tree are associated with the predefined nursing operation templates constructed using the following method: Define the root node of the rule tree as the nursing decision trigger event; In each non-leaf node of the rule tree, a judgment condition is set. The judgment condition is the result of the logical AND operation between the out-of-limit state of a specific combination of vital sign data and the code of a specific surgical stage. Starting from the root node, based on the excess status of vital signs data in the phased vital sign status snapshot and the current surgical stage code, the branch path that meets the conditions is traversed downwards layer by layer. When traversing to a leaf node, the leaf node is bound to a pre-configured nursing operation template for a specific combination of abnormal signs and a specific surgical stage scenario represented by the branch path. The nursing operation template includes a framework of nursing operation types to be performed, operation object identifiers, and parameter setting instructions.

Images