AI-based major trauma treatment decision support method and system

Abstract

This invention provides an AI-based decision support method and system for the treatment of severe trauma, relating to the field of artificial intelligence technology. The method includes acquiring the patient's physiological signs, injury location information, and geographical location; simultaneously acquiring the departmental resources, medical staff configuration, and equipment occupancy of medical institutions within the region; calculating quantitative indicators of injury urgency through multi-dimensional rules such as respiratory and circulatory stability assessment and bleeding risk assessment to classify patient priority levels; calculating the reachability time to medical institutions based on geographical coordinates to assess immediate reception capacity; generating an optimal medical institution allocation plan using a dual-objective optimization approach with time sensitivity constraints and resource matching constraints to ensure that the transfer and waiting time for high-priority patients is minimized and that medical resources are accurately matched to the patient's injury; and sending the allocation plan to the emergency dispatch system to trigger vehicle dispatch and resource reservation instructions.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an AI-based decision support method and system for the treatment of major trauma. Background Technology

[0002] Major trauma is one of the leading causes of death among young adults worldwide, and its treatment outcomes are closely related to the speed of emergency response and the efficiency of medical resource allocation. In existing trauma care systems, emergency dispatchers typically rely on telephone communication to obtain basic information about the injured, then assess the severity of the injuries based on experience, and select transport destinations based on the geographical location of medical institutions. This traditional model is inherently subjective; dispatchers struggle to accurately assess the true severity of a patient's condition in a short time, especially in cases of multiple or combined injuries, where verbal descriptions often fail to provide a comprehensive understanding of key physiological indicators such as respiratory and circulatory status and bleeding risk. Furthermore, the selection of medical institutions is primarily based on distance or established cooperative relationships, lacking dynamic monitoring of the target hospital's real-time resource status. This results in some critically ill patients being transferred to medical institutions with insufficient capacity or saturated resources, delaying optimal treatment.

[0003] A more prominent problem is the lack of a systematic resource matching mechanism in the existing treatment decision-making process. In many cases, patients arrive at the hospital only to find that the specialist they need is in surgery, critical equipment is occupied, or emergency beds are full, necessitating temporary coordination or a second transfer. This resource mismatch not only wastes precious treatment time but may also lead to a deterioration of the patient's condition due to waiting in the hospital. Traditional dispatching models cannot achieve precise matching between the patient's injury characteristics and the medical institution's treatment capabilities, nor can they reserve resources at the target hospital before transfer, resulting in significant efficiency bottlenecks in the entire treatment chain. With the increasingly complex distribution of urban medical resources, the coexistence of multiple medical institutions with significant differences in specialist capabilities, this decision-making method relying on human experience is no longer sufficient to meet the timeliness and accuracy requirements of major trauma treatment. Summary of the Invention

[0004] This invention provides an AI-based decision support method and system for the treatment of major trauma, which can solve the problems in the prior art.

[0005] A first aspect of this invention provides an AI-based decision support method for the treatment of major trauma, comprising: The system acquires vital signs data, injury site information, and current geographical location of trauma patients, as well as the status of departmental resources, staffing, and equipment occupancy of medical institutions within the region. Based on the physiological signs and injury site information, quantitative indicators of the patient's injury severity are calculated using multi-dimensional condition assessment rules, which include respiratory and circulatory stability assessment and bleeding risk assessment. Patients are then classified into different priority levels according to these quantitative indicators of injury severity. Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time of each medical institution is calculated. Combining the status of departmental resources, staffing, and equipment occupancy, the immediate reception capacity of each medical institution is determined. Based on the priority level, access time, and immediate reception capacity, an optimal medical institution allocation plan for the patient is generated through dual-objective optimization of time sensitivity constraints and resource matching constraints. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time for high-priority patients is minimized, and the resource matching constraint ensures that the departmental resources of the medical institutions are matched with the patient's injury site and quantifiable indicators of the severity of the injury. The optimal medical institution allocation plan is sent to the emergency dispatch system, triggering emergency vehicle dispatch instructions and resource reservation instructions for the target medical institution, thereby realizing the targeted transfer of the patient and the advance preparation of medical resources.

[0006] Based on the aforementioned physiological signs and injury site information, a multi-dimensional condition assessment rule is used to calculate quantifiable indicators of the patient's injury severity. Patients are then categorized into different priority levels based on these indicators, including: The physiological data were standardized, and features of heart rate variability, blood pressure fluctuation amplitude, and blood oxygen saturation decline rate were extracted to construct a physiological stability feature vector. Based on the information about the damaged site, the type and anatomical location of the damaged tissue are determined, and the severity score of the injury is calculated by combining the preset organ importance weighting coefficient and the injury depth grading rules. For the respiratory and circulatory stability assessment, the risk status of circulatory failure is identified by analyzing the correlation between the heart rate variability characteristics and the blood pressure fluctuation amplitude characteristics. For the bleeding risk assessment, the risk status of active bleeding is identified by analyzing the vascular distribution density corresponding to the injury site information and the blood pressure fluctuation amplitude characteristics. The physiological stability feature vector, injury severity score, circulatory failure risk status and active bleeding risk status are input into the preset injury urgency calculation rules to obtain the injury urgency quantification index. Based on the quantitative indicators of the severity of the injury and the preset priority classification threshold set, the patients are classified into priority levels of critical, emergency, secondary emergency, and delayed.

[0007] Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time to each medical institution is calculated. Combining this with the status of departmental resources, staffing levels, and equipment occupancy, the immediate reception capacity of each medical institution is determined, including: Based on the patient's current geographical location and the geographical coordinates of each medical institution, the theoretical driving distance to each medical institution is calculated. Combined with the current road traffic flow data and historical average vehicle speed data, the actual reachability time of each medical institution is corrected to obtain the actual reachability time. Analyze the resource status of the departments reported by each medical institution, and extract the number of vacant beds, available operating rooms, and vacant intensive care units for trauma surgery, orthopedics, neurosurgery, and interventional radiology. Analyze the staffing status of each medical institution, count the number of medical staff currently on duty, and determine the staffing sufficiency level of each department based on the preset staffing ratio standard; Analyze the equipment occupancy status of each medical institution to identify the real-time occupancy status and scheduled occupancy time of the medical equipment; The real-time patient reception capacity includes the number of vacant beds, the number of available operating rooms, the number of vacant intensive care units, the staff adequacy level, and the real-time occupancy status of equipment.

[0008] Based on the aforementioned priority level, access time, and immediate access capacity, before generating the optimal medical institution allocation plan for the patient through bi-objective optimization with time sensitivity constraints and resource matching constraints, the method further includes: A time-sensitive objective function is constructed, which is based on the time weight coefficient corresponding to the priority level, and the reachable time and the estimated in-hospital waiting time are weighted and summed. The estimated in-hospital waiting time is calculated based on the current number of patients queuing and the average processing time of the medical institution. A resource matching objective function is constructed. The resource matching objective function determines the set of required departmental resource types based on the patient's injury site information, calculates the matching degree between the medical institution's departmental resource status and the required set of departmental resource types, and combines the degree of fit between the quantification of injury urgency and the immediate reception capability score.

[0009] Based on the aforementioned priority level, availability, and immediate accessibility, an optimal medical institution allocation scheme for the patient is generated through dual-objective optimization constrained by time sensitivity and resource matching, including: The time-sensitive objective function and the resource matching objective function are normalized, and a dual-objective weight allocation strategy is determined according to the priority level. When the priority level is critical or urgent, the weight of the time-sensitive objective function is higher than the weight of the resource matching objective function. Using the Pareto optimal solution method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; The solution with the highest comprehensive score is selected from the Pareto optimal solution set as the optimal medical institution allocation scheme. The comprehensive score is obtained by weighting the time-sensitive objective function value and the resource matching degree objective function value according to the dual-objective weight allocation strategy.

[0010] Using the Pareto optimality method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; the solution with the highest comprehensive score from the Pareto optimal solution set is selected as the optimal medical institution allocation scheme, including: The constraints on the medical institution's capacity to receive patients are constructed, requiring that the candidate medical institution's immediate capacity score be no lower than the minimum capacity requirement corresponding to the quantitative indicator of the severity of the injury. The patient transport safety constraints are constructed, and the maximum allowable transport time is determined according to the priority level, requiring that the reachability time of the candidate medical institution does not exceed the maximum allowable transport time; In the set of medical institutions that meet the constraints of the medical institution's capacity to receive patients and the constraints of the patient transfer safety, the time sensitivity objective function value and the resource matching degree objective function value are calculated for each candidate medical institution, and a dual objective function value matrix is ​​constructed. Based on the dual objective function value matrix, Pareto dominance is determined. When a candidate medical institution is not inferior to other candidate medical institutions in both the time sensitivity objective function value and the resource matching objective function value, and is superior to other candidate medical institutions in at least one objective function value, the candidate medical institution is marked as a non-dominated solution. Extract all the non-dominated solutions to form the Pareto optimal solution set; Based on the aforementioned dual-objective weight allocation strategy, the time-sensitivity objective function value and the resource matching degree objective function value of each solution in the Pareto optimal solution set are weighted and summed to obtain the comprehensive score of each solution; The medical institution corresponding to the solution with the highest comprehensive score is selected as the optimal medical institution allocation scheme.

[0011] A second aspect of this invention provides an AI-based decision support system for the treatment of severe trauma, comprising: The data acquisition unit is used to acquire the physiological signs data, injury site information and current geographical location of trauma patients, and at the same time acquire the departmental resource status, medical staff configuration status and equipment occupancy status of various medical institutions in the area. The injury assessment unit is used to calculate the patient's injury urgency quantification index based on the physiological signs data and injury site information, and to use multi-dimensional condition assessment rules, including respiratory and circulatory stability assessment and bleeding risk assessment. Patients are classified into different priority levels according to the injury urgency quantification index. The capacity calculation unit is used to calculate the reachability time of each medical institution based on the patient's current geographical location and the geographical coordinates of each medical institution, and to determine the immediate reception capacity of each medical institution by combining the department's resource status, medical staff configuration, and equipment occupancy. The scheme optimization unit is used to generate the optimal medical institution allocation scheme for the patient based on the priority level, reachability and immediate reception capacity through dual-objective optimization of time sensitivity constraint and resource matching degree constraint. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time of high priority patients is minimized, and the resource matching degree constraint ensures that the departmental resources of the medical institution are matched with the patient's injury site and quantifiable indicators of the severity of injury. The dispatch execution unit is used to send the optimal medical institution allocation plan to the emergency dispatch system, trigger the emergency vehicle dispatch instruction and the resource reservation instruction of the target medical institution, so as to realize the targeted transfer of the patient and the advance preparation of medical resources.

[0012] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0014] The beneficial effects of this application are: This method addresses the problem of delayed treatment caused by information lag in traditional emergency dispatch by establishing a real-time correlation mechanism between patient physiological signs, injury sites, and medical resource status. The multi-dimensional condition assessment rules, combined with quantitative analysis of respiratory and circulatory stability and bleeding risk, can accurately identify the degree of life-threatening severity in patients with severe trauma, avoiding the risk of misjudgment caused by single-indicator assessments and making priority allocation more clinically targeted.

[0015] The dynamic integration of reachability calculation and immediate accessibility overcomes the limitations of traditional proximity-based allocation models. This method not only considers geographical distance but also incorporates departmental resource utilization, medical staff allocation, and equipment availability into the decision-making process, effectively mitigating the risk of secondary transfers due to overloaded medical institutions and significantly improving the success rate of treating critically ill patients.

[0016] The dual-objective optimization mechanism achieves a balanced decision between time sensitivity and resource matching. The time sensitivity constraint minimizes the sum of transport time and in-hospital waiting time, ensuring critically ill patients receive treatment within the golden window of opportunity. The resource matching constraint ensures precise matching of specialist resources with patient conditions, avoiding reduced treatment efficiency due to mismatches in professional capabilities. This collaborative optimization mechanism shortens pre-hospital delays while guaranteeing the immediate availability of in-hospital treatment resources. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the AI-based decision support method for major trauma treatment according to an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0020] Figure 1 This is a flowchart illustrating the AI-based decision support method for major trauma treatment according to an embodiment of the present invention. Figure 1 As shown, the method includes: The system acquires vital signs data, injury site information, and current geographical location of trauma patients, as well as the status of departmental resources, staffing, and equipment occupancy of medical institutions within the region. Based on the physiological signs and injury site information, quantitative indicators of the patient's injury severity are calculated using multi-dimensional condition assessment rules, which include respiratory and circulatory stability assessment and bleeding risk assessment. Patients are then classified into different priority levels according to these quantitative indicators of injury severity. Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time of each medical institution is calculated. Combining the status of departmental resources, staffing, and equipment occupancy, the immediate reception capacity of each medical institution is determined. Based on the priority level, access time, and immediate reception capacity, an optimal medical institution allocation plan for the patient is generated through dual-objective optimization of time sensitivity constraints and resource matching constraints. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time for high-priority patients is minimized, and the resource matching constraint ensures that the departmental resources of the medical institutions are matched with the patient's injury site and quantifiable indicators of the severity of the injury. The optimal medical institution allocation plan is sent to the emergency dispatch system, triggering emergency vehicle dispatch instructions and resource reservation instructions for the target medical institution, thereby realizing the targeted transfer of the patient and the advance preparation of medical resources.

[0021] Based on the aforementioned physiological signs and injury site information, a multi-dimensional condition assessment rule is used to calculate quantifiable indicators of the patient's injury severity. Patients are then categorized into different priority levels based on these indicators, including: The physiological data were standardized, and features of heart rate variability, blood pressure fluctuation amplitude, and blood oxygen saturation decline rate were extracted to construct a physiological stability feature vector. Based on the information about the damaged site, the type and anatomical location of the damaged tissue are determined, and the severity score of the injury is calculated by combining the preset organ importance weighting coefficient and the injury depth grading rules. For the respiratory and circulatory stability assessment, the risk status of circulatory failure is identified by analyzing the correlation between the heart rate variability characteristics and the blood pressure fluctuation amplitude characteristics. For the bleeding risk assessment, the risk status of active bleeding is identified by analyzing the vascular distribution density corresponding to the injury site information and the blood pressure fluctuation amplitude characteristics. The physiological stability feature vector, injury severity score, circulatory failure risk status and active bleeding risk status are input into the preset injury urgency calculation rules to obtain the injury urgency quantification index. Based on the quantitative indicators of the severity of the injury and the preset priority classification threshold set, the patients are classified into priority levels of critical, emergency, secondary emergency, and delayed.

[0022] During the disease assessment phase, the collected physiological data were standardized to eliminate dimensional differences. Specifically, heart rate values ​​were mapped to the 0-1 interval, and a min-maximum normalization method was used. The formula for calculating the normalized value was: the actual heart rate minus the lower limit of the normal heart rate, then divided by the normal heart rate range. When extracting heart rate variability features, the standard deviation of the continuous heartbeat interval was calculated. This indicator reflects the regulatory capacity of the autonomic nervous system; a value below 20 milliseconds suggests excessive sympathetic nerve activation. Blood pressure fluctuation amplitude features were obtained by calculating the range of continuous systolic blood pressure measurements; a range exceeding 30 mmHg was marked as significant fluctuation. The rate of decrease in blood oxygen saturation was analyzed using time series analysis, calculating the rate of change in saturation per unit time; a rate of decrease exceeding 2% per minute was marked as a rapid deterioration trend. These three types of features were combined to form a six-dimensional physiological stability feature vector.

[0023] The type of damaged tissue is determined based on the location of the injury; for example, liver injury is classified as a solid organ injury, and rib fractures are classified as skeletal system injuries. An organ importance weighting coefficient is introduced: 1.0 for the heart, lungs, and brain; 0.8 for the liver and spleen; and 0.3 for the limb bones. The injury depth grading rule uses a three-level classification: superficial injuries are assigned 1 point, partial-depth injuries are assigned 3 points, and full-depth or penetrating injuries are assigned 5 points. The injury severity score is equal to the product of the organ importance weighting coefficient and the injury depth grading score; when multiple injuries exist, the scores for each site are summed.

[0024] Respiratory and circulatory stability assessment was achieved using a decision tree model. Combinations of heart rate variability below a threshold and blood pressure fluctuation above a threshold were identified as circulatory failure risk states. Specifically, the judgment logic was as follows: when the standard deviation of heart rate variability was below 15 milliseconds, and the systolic blood pressure fluctuation exceeded 40 mmHg, or when the mean arterial pressure showed a continuous downward trend with a rate of decrease exceeding 5 mmHg per minute, a high-risk circulatory failure marker was output. Bleeding risk assessment was performed using an anatomical knowledge base. Based on the injury site information, a pre-stored vascular density map was queried, with the vascular density coefficients of the abdominal organ region and neck region set as high-risk levels. When the vascular density coefficient at the injury site was greater than 0.7, and the blood pressure fluctuation characteristics showed a stepwise decrease in systolic blood pressure, an active bleeding risk state was identified.

[0025] The calculation of injury severity uses a weighted summation method. The six components of the physiological stability feature vector are assigned weights of 0.15, 0.20, 0.15, 0.10, 0.20, and 0.20 respectively, and the weighted sum is used to obtain the physiological instability score. The injury severity score is directly used as the structural injury score. A risk correction value of 15 points is added when the risk of circulatory failure is true, and an additional 20 points is added when the risk of active bleeding is true. The quantitative index for injury severity equals the physiological instability score plus the injury severity score plus each risk correction value, with a value range of 0 to 100.

[0026] Priority is determined using a four-tiered threshold system. Injuries measured by a criticality index of 75 or higher are classified as critical, requiring immediate activation of the green channel. Indicators between 50 and 74 are classified as urgent, requiring transfer within the critical hour. Indicators between 25 and 49 are classified as secondary urgent, allowing for a transfer time extended to two hours. Indicators below 25 are classified as delayed, with transfer order determined based on overall medical resource availability.

[0027] Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time to each medical institution is calculated. Combining this with the status of departmental resources, staffing levels, and equipment occupancy, the immediate reception capacity of each medical institution is determined, including: Based on the patient's current geographical location and the geographical coordinates of each medical institution, the theoretical driving distance to each medical institution is calculated. Combined with the current road traffic flow data and historical average vehicle speed data, the actual reachability time of each medical institution is corrected to obtain the actual reachability time. Analyze the resource status of the departments reported by each medical institution, and extract the number of vacant beds, available operating rooms, and vacant intensive care units for trauma surgery, orthopedics, neurosurgery, and interventional radiology. Analyze the staffing status of each medical institution, count the number of medical staff currently on duty, and determine the staffing sufficiency level of each department based on the preset staffing ratio standard; Analyze the equipment occupancy status of each medical institution to identify the real-time occupancy status and scheduled occupancy time of the medical equipment; The real-time patient reception capacity includes the number of vacant beds, the number of available operating rooms, the number of vacant intensive care units, the staff adequacy level, and the real-time occupancy status of equipment.

[0028] To determine the immediate treatment capacity of medical institutions, the latitude and longitude coordinates of the trauma patient are first obtained from the positioning device. Simultaneously, the geographical coordinates of all medical institutions within the region with trauma treatment capabilities are retrieved from the medical institution database. The Haversine formula is used to calculate the theoretical straight-line distance between the patient's location and each medical institution, and then the distance is corrected based on the actual paths of the urban road network. The theoretical driving distance is obtained by multiplying the straight-line distance by a road tortuosity coefficient, which is preset according to the urban road network density and typically ranges from 1.3 to 1.5.

[0029] To obtain accurate reachability time, the system connects to the real-time traffic information interface of the traffic management department to obtain traffic flow data for each road segment during the current time period. Traffic flow data is presented as vehicle density, measured in vehicles per kilometer. Simultaneously, the system retrieves the average driving speed for the same time period over the past three months to establish a time-speed comparison table. When the real-time traffic flow exceeds 120% of the historical average, the estimated speed for the corresponding road segment is reduced by 30%; when the traffic flow is below 80% of the historical average, the speed is increased by 15%. The corrected speeds for each road segment are weighted and averaged to obtain the estimated driving speed. The actual reachability time is calculated by dividing the theoretical travel distance by this speed.

[0030] For analyzing departmental resource status, the system obtains real-time resource occupancy data for each department through the hospital information system interface. The number of vacant beds in the trauma surgery department is calculated through the bed management module, excluding admitted patients and reserved beds to obtain the number of available beds. The number of available operating rooms requires both no surgeries being performed and post-operative cleaning being completed; the system reads the surgery schedule and cleaning record to identify the number of operating rooms immediately available. The calculation of vacant intensive care unit beds takes into account equipment availability; only beds equipped with necessary equipment such as ventilators and monitors, and without patients occupying them, are counted as vacant. The resource status of orthopedics, neurosurgery, and interventional radiology departments is calculated using the same logic.

[0031] The analysis of medical staff allocation is completed through the human resources management system. The system captures the attendance data of the current shift and counts the actual number of attending physicians, resident physicians, and nurses on duty in each department. According to the staffing standards set by the health administration department, each open bed in the trauma surgery department requires 0.4 physicians and 1.2 nurses, and each operating room requires 2 physicians and 3 nurses. The actual number of staff on duty is compared with the standard requirements. When the actual number reaches more than 100% of the standard requirements, the staffing adequacy level is marked as "adequate"; when it reaches 80% to 100%, it is marked as "basically satisfied"; and when it is less than 80%, it is marked as "stressed".

[0032] Equipment occupancy identification involves large medical equipment such as CT, MRI, and DSA. The system connects to the equipment management platform and reads the equipment's operating status identifier, including three states: "Idle," "Under Inspection," and "Under Repair." For equipment in the "Under Inspection" state, the estimated end time of the current inspection is obtained. Simultaneously, the system retrieves the scheduling table from the equipment reservation system to identify scheduled inspection slots within the next two hours. Combining the real-time status with the scheduled slots, the system determines the equipment's availability at the patient's expected arrival time and generates a real-time equipment occupancy report.

[0033] The above analysis results are summarized to form a dataset assessing the real-time patient reception capacity of various medical institutions. This dataset includes three numerical indicators: the number of vacant beds, the number of available operating rooms, and the number of vacant intensive care units; a categorical indicator: staff adequacy level; and a composite indicator: real-time equipment occupancy status. Each indicator is accompanied by a timestamp to ensure the timeliness of the data. The system organizes the data according to the medical institution dimension, providing a quantitative basis for subsequent optimization and matching.

[0034] Based on the aforementioned priority level, access time, and immediate access capacity, before generating the optimal medical institution allocation plan for the patient through bi-objective optimization with time sensitivity constraints and resource matching constraints, the method further includes: A time-sensitive objective function is constructed, which is based on the time weight coefficient corresponding to the priority level, and the reachable time and the estimated in-hospital waiting time are weighted and summed. The estimated in-hospital waiting time is calculated based on the current number of patients queuing and the average processing time of the medical institution. A resource matching objective function is constructed. The resource matching objective function determines the set of required departmental resource types based on the patient's injury site information, calculates the matching degree between the medical institution's departmental resource status and the required set of departmental resource types, and combines the degree of fit between the quantification of injury urgency and the immediate reception capability score.

[0035] Before generating the optimal medical institution allocation plan, the two core objective functions of the bi-objective optimization need to be precisely constructed. The construction of the time-sensitive objective function first requires determining the corresponding time weight coefficient based on the patient's priority level. For patients classified as Level 1 priority, the time weight coefficient is set to a value between 0.7 and 0.9, reflecting the dominant role of transfer time in the overall objective. For Level 2 priority patients, the time weight coefficient is reduced to 0.4 to 0.6; and for Level 3 priority patients, the time weight coefficient is further reduced to 0.2 to 0.3. After obtaining the patient's current geographical location and the geographical coordinates of the target medical institution, the reachable time is calculated using real-time traffic data. This reachable time includes route planning time and actual travel time. The calculation of the estimated in-hospital waiting time requires obtaining the current number of patients in the medical institution's queue. This number is obtained through real-time data on patients awaiting treatment transmitted from the medical institution's information system. Simultaneously, the average processing time for similar trauma patients at the medical institution over the past seven days is obtained. Multiplying the current queue number by the average processing time yields the estimated in-hospital waiting time. The time-sensitive objective function is expressed as: the product of the time weight coefficient and the reachability time, plus the product of the supplementary weight coefficient and the estimated in-hospital waiting time. The supplementary weight coefficient is the difference between 1 and the time weight coefficient, ensuring that the sum of the two weights is 1.

[0036] In constructing the objective function for resource matching, the first step is to determine the set of required departmental resource types based on the patient's injury location information. When the injury location includes craniocerebral injury, the required departmental resource type set includes neurosurgery; when there is blunt trauma to the chest or abdomen, thoracic surgery or general surgery is added to the set; and when a fracture occurs, orthopedics is added. The matching degree between the medical institution's departmental resource status and the required departmental resource type set is calculated. The matching degree score is obtained by dividing the number of elements in the intersection of the sets by the number of elements in the required departmental resource type set, with a score ranging from 0 to 1. When a medical institution fully possesses all the departmental resource types required by the patient, the matching degree score is 1; when it only possesses some departmental resources, the score is reduced proportionally. The degree of fit is calculated by combining the quantifiable indicators of injury urgency with the immediate medical attention capacity score. When the quantifiable indicators of injury urgency are higher than 8, the immediate medical attention capacity score must be at least 7 to be considered a good match. A threshold judgment function is set; when the immediate medical attention capacity score meets the threshold requirement, the matching coefficient is 1; otherwise, the matching coefficient is 0.5. The resource matching objective function is expressed as the product of the department's resource matching score and its fit coefficient. This product serves as a quantifiable measure of resource matching for a single medical institution; the closer the value is to 1, the higher the degree of resource matching. In the dual-objective optimization process, the time-sensitive objective function is minimized, and the resource matching objective function is maximized. A non-dominated solution set is obtained using the Pareto front method. The optimal solution, which combines the values ​​of the time-sensitive objective function and the resource matching objective function, is selected as the final output.

[0037] Based on the aforementioned priority level, availability, and immediate accessibility, an optimal medical institution allocation scheme for the patient is generated through dual-objective optimization constrained by time sensitivity and resource matching, including: The time-sensitive objective function and the resource matching objective function are normalized, and a dual-objective weight allocation strategy is determined according to the priority level. When the priority level is critical or urgent, the weight of the time-sensitive objective function is higher than the weight of the resource matching objective function. Using the Pareto optimal solution method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; The solution with the highest comprehensive score is selected from the Pareto optimal solution set as the optimal medical institution allocation scheme. The comprehensive score is obtained by weighting the time-sensitive objective function value and the resource matching degree objective function value according to the dual-objective weight allocation strategy.

[0038] After determining patient priority levels, calculating the time to access medical institutions, and assessing immediate patient reception capacity, an optimal medical institution allocation plan for each patient needs to be generated through a dual-objective optimization mechanism.

[0039] The time-sensitive objective function and the resource matching objective function are normalized. The time-sensitive objective function is represented as the sum of transfer time and in-hospital waiting time. Transfer time is calculated based on geographical distance and traffic conditions, while in-hospital waiting time is estimated based on the current number of patients waiting in the target medical institution's emergency department and the average treatment time. The resource matching objective function represents the degree of matching between the medical institution's departmental resources and the patient's injury type. Specifically, it is calculated by comparing the specialist needs corresponding to the patient's injury site with the departmental configuration of the target medical institution. Since the two objective functions have different numerical dimensions, a maximum-minimum normalization method is used to map the values ​​of both objective functions to the interval between 0 and 1. A smaller normalized time-sensitive function value indicates a more significant time advantage, while a larger normalized resource matching function value indicates stronger resource suitability.

[0040] The dual-objective weighting strategy is determined based on the patient's priority level. When the priority level is critical, the weight of the time-sensitive objective function is set to 0.8, and the weight of the resource matching objective function is set to 0.2. This configuration ensures that critically ill patients are preferentially transferred to the nearest medical institution with basic treatment capabilities. When the priority level is emergency, the weight of the time-sensitive objective function is set to 0.7, and the weight of the resource matching objective function is set to 0.3. When the priority level is moderate, the weight of the time-sensitive objective function decreases to 0.4, and the weight of the resource matching objective function increases to 0.6. In this case, the focus is more on transferring the patient to a medical institution with sufficient departmental resources and the appropriate specialty.

[0041] A candidate solution set is generated using the Pareto optimality method. All medical institutions within the region are considered as alternatives, and a set of constraints is constructed, including constraints on the immediate reception capacity of medical institutions (the number of available emergency department beds in the target medical institution must be greater than zero, and the corresponding specialist physicians must be on duty); and constraints on patient transport safety (the transport time must not exceed the maximum tolerable transport time calculated based on quantifiable indicators of the patient's critical condition). Under these constraints, a fast non-dominated sorting genetic algorithm is used to solve the biobjective optimization problem, generating a Pareto optimal solution set. Each solution in this set represents a feasible medical institution allocation scheme, and there is no absolute dominance relationship between any two solutions.

[0042] The solution with the highest overall score from the Pareto optimal solution set is selected as the optimal medical institution allocation scheme. For each candidate scheme in the solution set, its normalized time-sensitivity objective function value and resource matching objective function value are multiplied by their corresponding weights, and then summed to obtain the overall score. The overall scores of all candidate schemes are compared, and the scheme with the highest score is selected as the final recommended medical institution allocation scheme. The target medical institution name, estimated transfer time, and receiving department information in this scheme are then encapsulated and output.

[0043] Using the Pareto optimal solution method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; Selecting the solution with the highest overall score from the Pareto optimal solution set as the optimal medical institution allocation scheme includes: The constraints on the medical institution's capacity to receive patients are constructed, requiring that the candidate medical institution's immediate capacity score be no lower than the minimum capacity requirement corresponding to the quantitative indicator of the severity of the injury. The patient transport safety constraints are constructed, and the maximum allowable transport time is determined according to the priority level, requiring that the reachability time of the candidate medical institution does not exceed the maximum allowable transport time; In the set of medical institutions that meet the constraints of the medical institution's capacity to receive patients and the constraints of the patient transfer safety, the time sensitivity objective function value and the resource matching degree objective function value are calculated for each candidate medical institution, and a dual objective function value matrix is ​​constructed. Based on the dual objective function value matrix, Pareto dominance is determined. When a candidate medical institution is not inferior to other candidate medical institutions in both the time sensitivity objective function value and the resource matching objective function value, and is superior to other candidate medical institutions in at least one objective function value, the candidate medical institution is marked as a non-dominated solution. Extract all the non-dominated solutions to form the Pareto optimal solution set; Based on the aforementioned dual-objective weight allocation strategy, the time-sensitivity objective function value and the resource matching degree objective function value of each solution in the Pareto optimal solution set are weighted and summed to obtain the comprehensive score of each solution; The medical institution corresponding to the solution with the highest comprehensive score is selected as the optimal medical institution allocation scheme.

[0044] The specific implementation process of the Pareto optimality solution method in generating the optimal allocation scheme for medical institutions is as follows: When constructing constraints on the patient reception capacity of medical institutions, the minimum reception capacity threshold is first determined based on quantifiable indicators of patient urgency. For example, if a patient's quantifiable indicator of patient urgency is 85 points, the corresponding minimum reception capacity requirement is set at 80 points. Candidate medical institutions are then screened, and their immediate reception capacity scores are calculated. The immediate reception capacity score comprehensively considers the number of trauma surgeons on duty, the number of available operating rooms, and the remaining intensive care unit (ICU) beds. Specifically, the number of on-duty trauma surgeons is multiplied by a weighting coefficient of 0.4, the number of available operating rooms by a weighting coefficient of 0.35, and the remaining ICU beds by a weighting coefficient of 0.25, and then summed. If a medical institution's immediate reception capacity score is below 80 points, it is removed from the candidate set.

[0045] When constructing safety constraints for patient transport, the maximum allowable transport time is set according to priority level. For Level 1 priority patients, the maximum allowable transport time is set at 15 minutes; for Level 2 priority patients, it is set at 25 minutes; and for Level 3 priority patients, it is set at 40 minutes. The reachability time of each candidate medical institution is compared with the maximum allowable transport time, and medical institutions whose reachability time exceeds the limit are excluded. The reachability time is calculated by comprehensively considering real-time traffic data and the average speed of emergency vehicles, using a shortest path algorithm between geographical coordinates.

[0046] Within the set of medical institutions satisfying the above two types of constraints, a dual objective function value is calculated for each candidate institution. The time-sensitivity objective function value is obtained by adding the reachability time to the estimated in-hospital waiting time, which is calculated based on the current emergency department queue length and average processing rate of the institution. The resource matching objective function value is obtained by weighted averaging of the resource completeness score, equipment availability score, and medical staff professional matching score corresponding to the injury site. For example, for patients with multiple fractures and visceral injuries, the weight for orthopedic resource completeness is set to 0.4, the weight for general surgery resource completeness is set to 0.35, and the weight for imaging equipment availability is set to 0.25.

[0047] Pareto dominance is determined based on the constructed dual objective function value matrix. The candidate medical institution set is traversed, and any two medical institutions are compared pairwise. If the time sensitivity objective function value of medical institution A is less than or equal to that of medical institution B, and the resource matching objective function value is greater than or equal to that of medical institution B, and at least one of its objective function values ​​is strictly superior to that of medical institution B, then medical institution A is determined to dominate medical institution B. Candidate medical institutions not dominated by any other medical institution are marked as non-dominated solutions; these non-dominated solutions together constitute the Pareto optimal solution set.

[0048] When selecting the optimal solution from the Pareto optimal solution set, a dual-objective weight allocation strategy is introduced. The weights are dynamically adjusted based on patient priority level. For first-priority patients, the weight for the time-sensitive objective is set to 0.7, and the weight for the resource matching objective is set to 0.3; for second-priority patients, the weights are set to 0.5 and 0.5 respectively. The two objective function values ​​for each medical institution in the Pareto optimal solution set are normalized: the time-sensitive objective function value is normalized using the reciprocal, and the resource matching objective function value is normalized using linear normalization. The normalized values ​​are then weighted and summed to obtain a comprehensive score. The medical institution with the highest comprehensive score is selected as the optimal allocation solution.

[0049] A second aspect of this invention provides an AI-based decision support system for the treatment of severe trauma, comprising: The data acquisition unit is used to acquire the physiological signs data, injury site information and current geographical location of trauma patients, and at the same time acquire the departmental resource status, medical staff configuration status and equipment occupancy status of various medical institutions in the area. The injury assessment unit is used to calculate the patient's injury urgency quantification index based on the physiological signs data and injury site information, and to use multi-dimensional condition assessment rules, including respiratory and circulatory stability assessment and bleeding risk assessment. Patients are classified into different priority levels according to the injury urgency quantification index. The capacity calculation unit is used to calculate the reachability time of each medical institution based on the patient's current geographical location and the geographical coordinates of each medical institution, and to determine the immediate reception capacity of each medical institution by combining the department's resource status, medical staff configuration, and equipment occupancy. The scheme optimization unit is used to generate the optimal medical institution allocation scheme for the patient based on the priority level, reachability and immediate reception capacity through dual-objective optimization of time sensitivity constraint and resource matching degree constraint. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time of high priority patients is minimized, and the resource matching degree constraint ensures that the departmental resources of the medical institution are matched with the patient's injury site and quantifiable indicators of the severity of injury. The dispatch execution unit is used to send the optimal medical institution allocation plan to the emergency dispatch system, trigger the emergency vehicle dispatch instruction and the resource reservation instruction of the target medical institution, so as to realize the targeted transfer of the patient and the advance preparation of medical resources.

[0050] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0051] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0052] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-based decision support method for the treatment of major trauma, characterized in that, include: The system acquires vital signs data, injury site information, and current geographical location of trauma patients, as well as the status of departmental resources, staffing, and equipment occupancy of medical institutions within the region. Based on the physiological signs and injury site information, quantitative indicators of the patient's injury severity are calculated using multi-dimensional condition assessment rules, which include respiratory and circulatory stability assessment and bleeding risk assessment. Patients are then classified into different priority levels according to these quantitative indicators of injury severity. Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time of each medical institution is calculated. Combining the status of departmental resources, staffing, and equipment occupancy, the immediate reception capacity of each medical institution is determined. Based on the priority level, access time, and immediate reception capacity, an optimal medical institution allocation plan for the patient is generated through dual-objective optimization of time sensitivity constraints and resource matching constraints. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time for high-priority patients is minimized, and the resource matching constraint ensures that the departmental resources of the medical institutions are matched with the patient's injury site and quantifiable indicators of the severity of the injury. The optimal medical institution allocation plan is sent to the emergency dispatch system, triggering emergency vehicle dispatch instructions and resource reservation instructions for the target medical institution, thereby realizing the targeted transfer of the patient and the advance preparation of medical resources.

2. The method according to claim 1, characterized in that, Based on the aforementioned physiological signs and injury site information, a multi-dimensional condition assessment rule is used to calculate quantifiable indicators of the patient's injury severity. Patients are then categorized into different priority levels based on these indicators, including: The physiological data were standardized, and features of heart rate variability, blood pressure fluctuation amplitude, and blood oxygen saturation decline rate were extracted to construct a physiological stability feature vector. Based on the information about the damaged site, the type and anatomical location of the damaged tissue are determined, and the severity score of the injury is calculated by combining the preset organ importance weighting coefficient and the injury depth grading rules. For the respiratory and circulatory stability assessment, the risk status of circulatory failure is identified by analyzing the correlation between the heart rate variability characteristics and the blood pressure fluctuation amplitude characteristics. For the bleeding risk assessment, the risk status of active bleeding is identified by analyzing the vascular distribution density corresponding to the injury site information and the blood pressure fluctuation amplitude characteristics. The physiological stability feature vector, injury severity score, circulatory failure risk status and active bleeding risk status are input into the preset injury urgency calculation rules to obtain the injury urgency quantification index. Based on the quantitative indicators of the severity of the injury and the preset priority classification threshold set, the patients are classified into priority levels of critical, emergency, secondary emergency, and delayed.

3. The method according to claim 1, characterized in that, Based on the patient's current geographical location and the geographical coordinates of each medical institution, the reachability time to each medical institution is calculated. Combining this with the status of departmental resources, staffing levels, and equipment occupancy, the immediate reception capacity of each medical institution is determined, including: Based on the patient's current geographical location and the geographical coordinates of each medical institution, the theoretical driving distance to each medical institution is calculated. Combined with the current road traffic flow data and historical average vehicle speed data, the actual reachability time of each medical institution is corrected to obtain the actual reachability time. Analyze the resource status of the departments reported by each medical institution, and extract the number of vacant beds, available operating rooms, and vacant intensive care units for trauma surgery, orthopedics, neurosurgery, and interventional radiology. Analyze the staffing status of each medical institution, count the number of medical staff currently on duty, and determine the staffing sufficiency level of each department based on the preset staffing ratio standard; Analyze the equipment occupancy status of each medical institution to identify the real-time occupancy status and scheduled occupancy time of the medical equipment; The real-time patient reception capacity includes the number of vacant beds, the number of available operating rooms, the number of vacant intensive care units, the staff adequacy level, and the real-time occupancy status of equipment.

4. The method according to claim 1, characterized in that, Based on the aforementioned priority level, access time, and immediate access capacity, before generating the optimal medical institution allocation plan for the patient through bi-objective optimization with time sensitivity constraints and resource matching constraints, the method further includes: A time-sensitive objective function is constructed, which is based on the time weight coefficient corresponding to the priority level, and the reachable time and the estimated in-hospital waiting time are weighted and summed. The estimated in-hospital waiting time is calculated based on the current number of patients queuing and the average processing time of the medical institution. A resource matching objective function is constructed. The resource matching objective function determines the set of required departmental resource types based on the patient's injury site information, calculates the matching degree between the medical institution's departmental resource status and the required set of departmental resource types, and combines the degree of fit between the quantification of injury urgency and the immediate reception capability score.

5. The method according to claim 4, characterized in that, Based on the aforementioned priority level, availability, and immediate accessibility, an optimal medical institution allocation scheme for the patient is generated through dual-objective optimization constrained by time sensitivity and resource matching, including: The time-sensitive objective function and the resource matching objective function are normalized, and a dual-objective weight allocation strategy is determined according to the priority level. When the priority level is critical or urgent, the weight of the time-sensitive objective function is higher than the weight of the resource matching objective function. Using the Pareto optimal solution method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; The solution with the highest comprehensive score is selected from the Pareto optimal solution set as the optimal medical institution allocation scheme. The comprehensive score is obtained by weighting the time-sensitive objective function value and the resource matching degree objective function value according to the dual-objective weight allocation strategy.

6. The method according to claim 5, characterized in that, Using the Pareto optimality method, under the premise of satisfying the constraints of the medical institution's capacity to receive patients and the patient transfer safety constraints, the Pareto optimal solution set of the time-sensitive objective function and the resource matching degree objective function is obtained; the solution with the highest comprehensive score from the Pareto optimal solution set is selected as the optimal medical institution allocation scheme, including: The constraints on the medical institution's capacity to receive patients are constructed, requiring that the candidate medical institution's immediate capacity score be no lower than the minimum capacity requirement corresponding to the quantitative indicator of the severity of the injury. The patient transport safety constraints are constructed, and the maximum allowable transport time is determined according to the priority level, requiring that the reachability time of the candidate medical institution does not exceed the maximum allowable transport time; In the set of medical institutions that meet the constraints of the medical institution's capacity to receive patients and the constraints of the patient transfer safety, the time sensitivity objective function value and the resource matching degree objective function value are calculated for each candidate medical institution, and a dual objective function value matrix is ​​constructed. Based on the dual objective function value matrix, Pareto dominance is determined. When a candidate medical institution is not inferior to other candidate medical institutions in both the time sensitivity objective function value and the resource matching objective function value, and is superior to other candidate medical institutions in at least one objective function value, the candidate medical institution is marked as a non-dominated solution. Extract all the non-dominated solutions to form the Pareto optimal solution set; Based on the aforementioned dual-objective weight allocation strategy, the time-sensitivity objective function value and the resource matching degree objective function value of each solution in the Pareto optimal solution set are weighted and summed to obtain the comprehensive score of each solution; The medical institution corresponding to the solution with the highest comprehensive score is selected as the optimal medical institution allocation scheme.

7. An AI-based decision support system for the treatment of severe trauma, used to implement the method as described in any one of claims 1-6, characterized in that, include: The data acquisition unit is used to acquire the physiological signs data, injury site information and current geographical location of trauma patients, and at the same time acquire the departmental resource status, medical staff configuration status and equipment occupancy status of various medical institutions in the area. The injury assessment unit is used to calculate the patient's injury urgency quantification index based on the physiological signs data and injury site information, and to use multi-dimensional condition assessment rules, including respiratory and circulatory stability assessment and bleeding risk assessment. Patients are classified into different priority levels according to the injury urgency quantification index. The capacity calculation unit is used to calculate the reachability time of each medical institution based on the patient's current geographical location and the geographical coordinates of each medical institution, and to determine the immediate reception capacity of each medical institution by combining the department's resource status, medical staff configuration, and equipment occupancy. The scheme optimization unit is used to generate the optimal medical institution allocation scheme for the patient based on the priority level, reachability and immediate reception capacity through dual-objective optimization of time sensitivity constraint and resource matching degree constraint. The time sensitivity constraint ensures that the sum of the transfer time and in-hospital waiting time of high priority patients is minimized, and the resource matching degree constraint ensures that the departmental resources of the medical institution are matched with the patient's injury site and quantifiable indicators of the severity of injury. The dispatch execution unit is used to send the optimal medical institution allocation plan to the emergency dispatch system, trigger the emergency vehicle dispatch instruction and the resource reservation instruction of the target medical institution, so as to realize the targeted transfer of the patient and the advance preparation of medical resources.

8. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 6.

Images