Regional major trauma data processing method and system based on cloud computing

Abstract

This invention provides a cloud-based method and system for processing regional major trauma data, relating to the field of cloud computing technology. It includes acquiring and processing raw data of trauma patients through a cloud computing platform, generating dynamic severity ratings, and combining this with a regional medical resource availability matrix. Cross-institutional collaborative matching and scheduling is performed under constraints of transport time, treatment matching degree, and resource balance, generating and issuing optimal scheduling instructions. This invention achieves dynamic and accurate assessment of trauma patient severity and efficient collaborative scheduling of regional medical resources.

Description

Technical Field

[0001] This invention relates to the field of cloud computing technology, and in particular to a cloud-based method and system for processing regional major trauma data. Background Technology

[0002] In the field of regional major trauma care, existing technologies typically rely on a centralized data processing model. Each medical institution independently collects physiological parameter data from trauma patients and conducts manual or semi-automatic severity assessments in its local system using fixed scoring scales, such as the Injury Severity Score (ISS) or the Brief Injury Classification (AIS). Subsequently, hospital dispatchers, based on their own institution's resources and limited external information, communicate via telephone or regional information platforms to decide whether to accept patients or conduct inter-hospital transfers. Throughout this process, data remains siloed, assessment standards and resource status are updated slowly, the decision-making process heavily relies on human experience, and there is a lack of real-time, holistic understanding of the overall regional resource situation.

[0003] This conventional approach has significant drawbacks. Firstly, static scoring mechanisms and discrete data processing cannot effectively address the dynamic evolution of trauma patients' conditions. Fixed-weight scoring rules struggle to capture the real-time fluctuations in physiological parameters reflecting changes in severity, leading to assessment results lagging behind clinical reality and potentially delaying the identification and priority treatment of patients whose conditions are rapidly deteriorating. Secondly, resource allocation decisions are severely fragmented and lack optimization. Allocation decisions are often limited to simple negotiations between two institutions, failing to comprehensively consider the needs of all patients awaiting treatment at the regional level, the real-time specialist capabilities of each institution, bed capacity, and multi-dimensional constraints such as geographical accessibility. This easily results in uneven distribution of treatment resources, with some institutions overburdened while others have idle resources, and the overall efficiency and fairness of treatment need improvement. Summary of the Invention

[0004] This invention provides a cloud computing-based method and system for processing regional major trauma data, which can solve the problems in the prior art.

[0005] A first aspect of the present invention provides a cloud computing-based method for processing regional major trauma data, comprising: The system acquires raw data of trauma patients reported by medical nodes within the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform. It generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. Based on the dynamic severity rating, the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix. Based on the dynamic severity rating and the regional medical resource availability matrix, and with patient transfer time, treatment capacity matching degree, and resource load balance as constraints, cross-institutional collaborative matching and scheduling operations are performed on trauma patients and medical resources to generate optimal cross-institutional scheduling instructions. The optimal cross-agency scheduling instruction is sent to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

[0006] Parallel distributed scoring calculations are performed on a cloud computing platform to generate an initial score for each patient based on trauma severity scoring rules. The weights of each dimension are then adaptively updated according to real-time changes in the initial score to obtain a dynamic severity rating for each trauma patient, including: The multidimensional physiological parameters are divided into multiple scoring sub-tasks according to the dimension category and distributed to multiple parallel computing nodes of the cloud computing platform. Each parallel computing node performs independent scoring calculations on the assigned scoring sub-tasks according to the trauma severity scoring rules to obtain the scores for each dimension. At the aggregation node, a weighted aggregation of the scores for each dimension is performed based on the current weight of each dimension to obtain the initial score for each patient; Based on the change in the initial score within a continuous evaluation period, the contribution of each scoring dimension to the change in the initial score is identified. For dimensions whose contribution exceeds the current weight representation range, the corresponding weight is increased; for dimensions whose contribution is less than the current weight representation range, the corresponding weight is decreased, thus completing the adaptive update of the scoring weights of each dimension. The scores of each dimension are re-weighted and aggregated according to the adaptively updated score weights to obtain an updated score. The updated score is then mapped to the corresponding severity level according to the severity grading mapping rule to obtain the dynamic severity rating for each trauma patient.

[0007] The real-time bed occupancy rate, specialty treatment capacity index, and geographical coordinates of each medical institution are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix, including: The current bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are obtained by polling the regional medical resource database according to the medical institution identifier. The available bed capacity of each medical institution is calculated based on the current bed occupancy rate. The degree of suitability of each medical institution for the treatment of various types of trauma is determined based on the specialty treatment capacity index. Using the medical institution identifier as the row index and the available bed capacity, the degree of adaptation of each specialty treatment, and the geographical coordinates as column attributes, the acquired medical institution resource data are written into the corresponding positions in the matrix to obtain the regional medical resource availability status matrix. When resource data in the regional medical resource database changes, a partial update is performed on the available bed capacity and specialty treatment suitability of the corresponding row in the regional medical resource availability status matrix, and a timestamp is added to the updated matrix elements to ensure that the regional medical resource availability status matrix is ​​consistent with the data status of the regional medical resource database, thus obtaining the updated regional medical resource availability status matrix.

[0008] When resource data in the regional medical resource database changes, a partial update is performed on the corresponding row of the regional medical resource availability status matrix, and a timestamp is appended, including: Continuously monitor data change events in the regional medical resource database, and extract the medical institution identifier, changed field type, and changed value carried in the change event when a change event is detected; Based on the medical institution identifier, locate the corresponding row in the regional medical resource availability status matrix, determine the column attribute to be updated based on the changed field type, write the changed value into the column attribute position to be updated in the corresponding row, and append the current time timestamp to that position to complete the partial update of the regional medical resource availability status matrix. Before the regional medical resource availability status matrix is ​​accessed by the scheduling operation, the time interval of each element from the current time is calculated based on the timestamp of each matrix element. For matrix elements whose time interval exceeds the preset validity period, the latest value is retrieved from the regional medical resource database and overwritten. The timestamp of the overwritten element is updated to ensure that each element in the regional medical resource availability status matrix used by the scheduling operation is within the valid time range, and the updated regional medical resource availability status matrix is ​​obtained.

[0009] Based on the dynamic severity rating and the regional medical resource availability matrix, and constrained by patient transfer time, treatment capacity matching degree, and resource load balance, a cross-institutional collaborative matching and scheduling operation is performed on trauma patients and medical resources to generate optimal cross-institutional scheduling instructions, including: Based on the dynamic severity rating, the urgency level of treatment and the type of specialist treatment required for each trauma patient are determined. Based on the available bed capacity and specialist treatment suitability of each medical institution in the regional medical resource availability status matrix, a set of candidate matching pairs is generated. For each matching pair in the candidate matching pair set, the estimated transfer time from the patient to the target medical institution is calculated by combining the geographical coordinates in the regional medical resource availability status matrix. The transfer time constraint is verified based on the estimated transfer time and the urgency level of the patient's treatment. Matching pairs that do not meet the transfer time constraint are eliminated, and candidate matching pairs that meet the transfer time constraint are obtained. Among the candidate matching pairs that meet the transfer time constraint, matching pairs whose specialty treatment suitability of the target medical institution does not meet the type of specialty treatment required by the patient are screened out according to the treatment capacity matching degree constraint. In the remaining candidate matching pairs, the balance of the number of patients carried by each medical institution is checked according to the resource load balance degree constraint. Matching combinations whose load difference exceeds the balance constraint are adjusted to generate the optimal cross-institution scheduling instruction that meets all constraints.

[0010] Based on resource load balance constraints, a balance check is performed on the number of patients carried by each medical institution. Matching combinations where load differences exceed the balance constraints are adjusted, and an optimal cross-institution scheduling instruction that satisfies all constraints is generated, including: The total number of patients assigned to each target medical institution in the remaining candidate matching pairs is counted. The total number of patients is divided by the available bed capacity of the corresponding medical institution to obtain the current load occupancy rate of each medical institution. All medical institutions in the region are then sorted according to their load occupancy rates. The difference between the maximum and minimum load occupancy rates of each medical institution in the region is calculated to obtain the load difference degree. The load difference degree is compared with the resource load balance constraint. For the matching combination where the load difference degree exceeds the balance constraint, the medical institution with the highest load occupancy rate and the transferable patients it carries are identified. The transferable patients are reassigned from the medical institution with the highest load difference to other medical institutions that meet the constraints of transfer time and treatment capacity matching degree and have the lowest current load occupancy rate. The load difference of each medical institution is recalculated for the reassigned matching combination. The above reassignment operation is repeated until the load difference of all medical institutions meets the balance constraint. All the finally determined matching pairs are encapsulated into the optimal cross-institution scheduling instruction.

[0011] The optimal cross-agency scheduling instruction is distributed to the corresponding medical node via the cloud computing platform. The corresponding medical node then executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction, including: The cloud computing platform, based on the medical node identifier in the optimal cross-institutional scheduling instruction, splits the optimal cross-institutional scheduling instruction into sub-instructions corresponding to each target medical node, sends each sub-instruction to the corresponding medical node, confirms the delivery status of each sub-instruction, and retransmits sub-instructions that do not return a confirmation response within a preset time limit. After receiving the sub-instruction, the corresponding medical node parses the patient information, estimated arrival time and required treatment resource type in the sub-instruction, issues a resource preparation instruction to the corresponding department based on the estimated arrival time, completes the preparatory deployment of the corresponding treatment resources to the designated reception area, and reports the deployment status of the treatment resources to the cloud computing platform. After aggregating the deployment status of medical resources reported by each medical node, the cloud computing platform identifies medical nodes with resource gaps and sends resource replenishment requests to other medical nodes in the region with available resources, thereby completing the cross-node collaborative allocation of trauma treatment resources in the region.

[0012] A second aspect of the present invention provides a cloud-based regional major trauma data processing system, comprising: The data acquisition unit is used to acquire the raw data of trauma patients reported by each medical node in the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform, generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. The resource status unit is used to extract the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution from the regional medical resource database of the cloud computing platform based on the dynamic severity rating, and to construct a regional medical resource availability status matrix. The collaborative matching unit is used to perform cross-institutional collaborative matching scheduling operations on trauma patients and medical resources based on the dynamic severity rating and the regional medical resource availability status matrix, with patient transfer time, treatment capacity matching degree and resource load balance as constraints, and generate the optimal cross-institutional scheduling instruction. The instruction issuing unit is used to issue the optimal cross-agency scheduling instruction to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

[0013] 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.

[0014] 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.

[0015] By constructing a regional medical resource availability matrix, this method integrates multi-dimensional dynamic information such as real-time bed occupancy rates, specialist treatment capabilities, and geographical coordinates, providing a comprehensive and accurate data foundation for resource allocation. This approach enables real-time perception and unified management of the global status of regional medical resources.

[0016] By using patient transfer time, matching degree of treatment capacity, and resource load balance as constraints, cross-institutional collaborative matching and scheduling calculations can quickly generate optimal scheduling instructions under complex constraints. This effectively shortens the transfer and waiting time for trauma patients, ensures rapid matching of patients with the most suitable treatment resources, and improves the timeliness of treatment. At the same time, the consideration of resource load balance avoids the overload of resources in a single medical institution, promotes the rational and balanced use of regional medical resources, and enhances the resilience and efficiency of the overall regional trauma treatment system.

[0017] The optimal scheduling instructions are directly sent to the corresponding medical nodes via the cloud computing platform, ensuring rapid and accurate instruction transmission. These medical nodes can then prepare for patient reception and allocate resources in advance, achieving seamless integration from decision-making to execution. This saves valuable time for trauma patients and optimizes the overall process and efficiency of regional major trauma emergency response. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the regional major trauma data processing method based on cloud computing, as described in an embodiment of the present invention. Detailed Implementation

[0019] 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.

[0020] 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.

[0021] Figure 1 This is a flowchart illustrating a cloud-based regional major trauma data processing method according to an embodiment of the present invention. Figure 1 As shown, the method includes: The system acquires raw data of trauma patients reported by medical nodes within the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform. It generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. Based on the dynamic severity rating, the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix. Based on the dynamic severity rating and the regional medical resource availability matrix, and with patient transfer time, treatment capacity matching degree, and resource load balance as constraints, cross-institutional collaborative matching and scheduling operations are performed on trauma patients and medical resources to generate optimal cross-institutional scheduling instructions. The optimal cross-agency scheduling instruction is sent to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

[0022] For example, parallel distributed scoring calculation is performed on a cloud computing platform to generate an initial score for each patient based on trauma severity scoring rules. The scoring weights for each dimension are then adaptively updated based on real-time changes in the initial score to obtain a dynamic severity rating for each trauma patient, including: The multidimensional physiological parameters are divided into multiple scoring sub-tasks according to the dimension category and distributed to multiple parallel computing nodes of the cloud computing platform. Each parallel computing node performs independent scoring calculations on the assigned scoring sub-tasks according to the trauma severity scoring rules to obtain the scores for each dimension. At the aggregation node, a weighted aggregation of the scores for each dimension is performed based on the current weight of each dimension to obtain the initial score for each patient; Based on the change in the initial score within a continuous evaluation period, the contribution of each scoring dimension to the change in the initial score is identified. For dimensions whose contribution exceeds the current weight representation range, the corresponding weight is increased; for dimensions whose contribution is less than the current weight representation range, the corresponding weight is decreased, thus completing the adaptive update of the scoring weights of each dimension. The scores of each dimension are re-weighted and aggregated according to the adaptively updated score weights to obtain an updated score. The updated score is then mapped to the corresponding severity level according to the severity grading mapping rule to obtain the dynamic severity rating for each trauma patient.

[0023] After obtaining the multidimensional physiological parameters of the trauma patient, the cloud computing platform initiates a parallel distributed scoring calculation process. The platform first breaks down the multidimensional physiological parameters according to their dimensional categories, organizing different categories of parameters such as physiological indicators, biochemical indicators, imaging characteristics, and vital signs into independent scoring subtasks. Each scoring subtask contains all parameter data for a specific dimension and the corresponding scoring rules. The cloud computing platform scheduler allocates these scoring subtasks to different parallel computing nodes based on the number of available computing nodes and the computational complexity of each scoring subtask. For example, scoring subtasks consisting of vital signs such as heart rate, blood pressure, and respiratory rate are allocated to computing node A; biochemical indicators such as hemoglobin, lactate levels, and coagulation function are allocated to computing node B; and imaging scoring subtasks such as CT values ​​and the number of fracture sites are allocated to computing node C.

[0024] After receiving the scoring subtask, each parallel computing node performs independent scoring calculations according to the trauma severity scoring rules. The trauma severity scoring rules define the normal range, abnormality grading standards, and corresponding scores for each dimension parameter. Taking vital signs scoring as an example, a baseline score of 0 is assigned when systolic blood pressure is between 90-120 mmHg, 2 points when it is between 70-89 mmHg, and 4 points when it is below 70 mmHg. The computing nodes read the patient's real-time systolic blood pressure value and match it with the interval thresholds in the scoring rules to determine the sub-score for that parameter. For heart rate parameters, a normal range of 60-100 beats per minute corresponds to 0 points, tachycardia of 100-120 beats per minute corresponds to 1 point, severe tachycardia exceeding 120 beats per minute corresponds to 3 points, and bradycardia below 60 beats per minute also corresponds to a corresponding score. Each computing node scores all parameters within its assigned dimension separately and then summarizes the scores within that dimension to obtain the sub-score for that dimension. The score for the vital signs dimension may be 8 points, the score for the biochemical indicators dimension may be 6 points, and the score for the imaging dimension may be 12 points.

[0025] After each parallel computing node completes its independent scoring calculation, it uploads the scores for each dimension to the aggregation node. The aggregation node maintains a current weight configuration table for each scoring dimension. The initial weights can be set based on clinical experience, for example, a weight of 0.35 for vital signs, 0.25 for biochemical indicators, 0.30 for imaging, and 0.10 for vital signs parameters. The aggregation node performs a weighted aggregation calculation, multiplying each dimension's score by its corresponding weight and summing the results to obtain the patient's initial score. Specifically, a vital signs score of 8 points multiplied by a weight of 0.35 equals 2.8 points; a biochemical indicators score of 6 points multiplied by a weight of 0.25 equals 1.5 points; an imaging score of 12 points multiplied by a weight of 0.30 equals 3.6 points; and a vital signs parameter score of 4 points multiplied by a weight of 0.10 equals 0.4 points. The sum of these four weighted scores yields an initial score of 8.3 points.

[0026] The system is set to a continuous assessment cycle of 10 minutes. At the end of each assessment cycle, a complete parallel scoring calculation is performed again to obtain a new initial score. The change in score is calculated by comparing the initial scores of adjacent assessment cycles. Assuming the initial score of the first cycle is 8.3 points and the initial score of the second cycle is 11.5 points, the change is 3.2 points. To identify the contribution of each scoring dimension to the change in the initial score, the system tracks the change in the scores of each dimension's sub-items over the two weeks. The score of the vital signs dimension sub-item increases from 8 points to 14 points, an increase of 6 points; the score of the biochemical indicators dimension sub-item increases from 6 points to 7 points, an increase of 1 point; the score of the imaging dimension sub-item increases from 12 points to 13 points, an increase of 1 point; and the score of the vital signs parameter dimension sub-item remains at 4 points, with no change.

[0027] The actual contribution of each dimension to the change in the total score was calculated. The vital signs dimension contributed 6 points multiplied by the current weight of 0.35, which equals 2.1 points. The biochemical indicators dimension contributed 1 point multiplied by 0.25, which equals 0.25 points. The imaging dimension contributed 1 point multiplied by 0.30, which equals 0.30 points. The vital signs parameter dimension contributed 0 points. The contribution level was compared with the representation range of the current weight. If the proportion of a dimension's contribution to the total change was significantly higher than the proportion of its current weight to the total weight, then the contribution of that dimension was determined to be outside the representation range of the current weight. The vital signs dimension contributed 2.1 points, accounting for 65.6% of the total change of 3.2 points, while its current weight of 0.35 accounted for only 35% of the total weight of 1.0. Therefore, the contribution of this dimension was determined to be outside the representation range, and its weight was adjusted upward from 0.35 to 0.42. Correspondingly, the contribution proportions of the biochemical indicators and imaging dimensions were lower than their respective weight proportions, and their weights were adjusted downward to 0.22 and 0.27, respectively. Since the vital signs dimension has no contribution and the smallest weight, its weight is kept constant at 0.09 to ensure that the sum of the weights of all dimensions is 1.0.

[0028] After the weight adaptive update is completed, the aggregation node uses the new weights to re-weight and aggregate the scores of each dimension in the current period. The vital signs score of 14 points multiplied by the updated weight of 0.42 equals 5.88 points, the biochemical indicators score of 7 points multiplied by 0.22 equals 1.54 points, the imaging score of 13 points multiplied by 0.27 equals 3.51 points, and the vital signs parameter score of 4 points multiplied by 0.09 equals 0.36 points. The sum of the four items gives an updated score of 11.29 points.

[0029] The system has pre-defined severity grading mapping rules that map continuous scores to discrete severity levels. The mapping rules define a score of 0-5 as mild trauma, 5-10 as moderate trauma, 10-15 as severe trauma, and above 15 as critical trauma. The aggregation node matches the updated score of 11.29 with the interval thresholds in the mapping rules to determine the patient's current severity level as severe trauma. This dynamic severity rating is updated in real time with each assessment cycle's score calculation and weight updates, accurately reflecting the evolution of the patient's condition and providing continuous quantitative evidence for clinical decision-making.

[0030] For example, real-time bed occupancy rates, specialty treatment capacity indices, and geographical coordinates of various medical institutions are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix, including: The current bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are obtained by polling the regional medical resource database according to the medical institution identifier. The available bed capacity of each medical institution is calculated based on the current bed occupancy rate. The degree of suitability of each medical institution for the treatment of various types of trauma is determined based on the specialty treatment capacity index. Using the medical institution identifier as the row index and the available bed capacity, the degree of adaptation of each specialty treatment, and the geographical coordinates as column attributes, the acquired medical institution resource data are written into the corresponding positions in the matrix to obtain the regional medical resource availability status matrix. When resource data in the regional medical resource database changes, a partial update is performed on the available bed capacity and specialty treatment suitability of the corresponding row in the regional medical resource availability status matrix, and a timestamp is added to the updated matrix elements to ensure that the regional medical resource availability status matrix is ​​consistent with the data status of the regional medical resource database, thus obtaining the updated regional medical resource availability status matrix.

[0031] In the dynamic management of regional medical resources, it is necessary to extract operational status data of each medical institution in real time from the regional medical resource database of the cloud computing platform. The regional medical resource database typically uses a distributed database architecture to store basic information and dynamic data of each medical institution, including medical institution identification, bed configuration information, department settings, medical staff allocation, equipment resource status, and geographical location information. For trauma treatment scenarios, the focus is on three core data types: bed occupancy rate, specialist treatment capacity index, and geographical coordinates.

[0032] Following the coding rules for medical institution identification, real-time data for each medical institution is retrieved one by one from the regional medical resource database. A medical institution identification typically consists of an administrative division code, an institution type code, and a sequence number. For example, "110101-H01-0023" represents the 23rd general hospital in a certain region. Using this identification as a query key, the current bed occupancy rate of that institution is extracted from the database. The bed occupancy rate is calculated based on the ratio of occupied beds to the total number of beds. For example, if a hospital has a total of 500 beds and 420 beds are occupied, then the current bed occupancy rate is 84%. Based on this occupancy rate, the available bed capacity can be calculated by multiplying the total number of beds by the percentage of remaining occupied space; therefore, the available bed capacity for this hospital is 80 beds.

[0033] The Specialty Treatment Capability Index is a quantitative indicator measuring a medical institution's ability to handle specific types of trauma. This index comprehensively considers factors such as the number of specialist physicians, operating room configuration, completeness of specialized equipment, and historical treatment success rates. During the data extraction phase, capability scores for each medical institution in trauma-related departments such as neurosurgery, orthopedics, thoracic surgery, and general surgery are retrieved from a resource database. For example, a tertiary hospital's neurosurgery department has 3 chief physicians and 5 associate chief physicians, possesses 2 dedicated neurosurgery operating rooms, 1 neuronavigation system, and complete microsurgical equipment, achieving a 92% success rate in treating severe traumatic brain injury over the past three years. Its neurosurgery treatment capability index is assessed at 0.95. A mapping relationship is established for different types of trauma, with a direct correspondence between traumatic brain injury and the neurosurgery capability index, and between pelvic fracture and orthopedic capability indices. This mapping determines the medical institution's suitability for treating various types of trauma.

[0034] Geographic coordinate data is extracted from the geographic information field in the resource repository, and the latitude and longitude coordinates of the emergency room entrance of the medical institution are recorded using the WGS-84 coordinate system. For example, the longitude of the emergency room entrance of a certain hospital is 116.3972°E and the latitude is 39.9087°N. Geographic coordinate data is the basic parameter for subsequent calculation of transfer distance and time.

[0035] After data extraction, a regional medical resource availability matrix is ​​constructed. The matrix uses a two-dimensional array structure, with row indices corresponding to the identifiers of each medical institution. Column attributes include available bed capacity, suitability for various specialties, and geographical coordinates. Assuming there are 50 medical institutions in the region, covering 6 major types of trauma, the matrix has 50 rows and 9 columns: 1 column for available bed capacity, 6 columns for suitability for specialties, and 2 columns for geographical coordinates. Taking the i-th medical institution as an example, its identifier is used as the row index. The i-th row of the matrix is ​​then filled with the available bed capacity, suitability for treatment of craniocerebral trauma, suitability for treatment of chest and abdominal trauma, suitability for treatment of pelvic fractures, suitability for treatment of spinal trauma, suitability for treatment of multiple injuries, suitability for treatment of burns, longitude coordinates, and latitude coordinates. Organizing data in matrix form efficiently supports batch queries and multi-dimensional filtering operations.

[0036] To maintain consistency between the matrix data and the resource database, an incremental update mechanism is established. When resource data in the regional medical resource database changes, the change event is captured through database triggers or message queues. The change event includes the medical institution identifier, change type, changed fields, and new values. Upon receiving the change event, the corresponding row in the matrix is ​​located based on the medical institution identifier. If the change involves bed occupancy, the available bed capacity of the institution is recalculated, and the first column of the corresponding row in the matrix is ​​updated. If the change involves specialist personnel or equipment configuration, the treatment capacity index of the relevant specialty is reassessed, and the specialty treatment suitability column of the corresponding row in the matrix is ​​updated. When performing the update operation, only the changed matrix elements are modified, rather than the entire matrix is ​​rebuilt. This local update method significantly reduces computational overhead.

[0037] Each time a matrix element is updated, a timestamp is appended to record the update time. The timestamp uses the Unix timestamp format, accurate to the second. For example, if a hospital admits 5 new patients at 14:30:25 on January 15, 2024, reducing its available bed capacity from 80 to 75, the timestamp 1705300225 would be appended when updating this matrix element. Timestamps allow tracking of changes in the resource status of various medical institutions, providing a temporal basis for data backtracking or anomaly investigation. Simultaneously, timestamps can also be used to determine the freshness of matrix data. When the difference between the timestamp of a matrix element and the current time exceeds a set threshold, an active refresh mechanism is triggered to retrieve the latest data from the resource repository, ensuring the timeliness of the matrix data.

[0038] Through continuous incremental updates and maintenance, a matrix reflecting the real-time availability of regional medical resources is obtained. This matrix serves as the data foundation for subsequent transfer decisions, supporting the rapid screening of medical institutions that meet treatment needs and have available resources, and providing data support for the precise scheduling of trauma treatment.

[0039] For example, when resource data in the regional medical resource database changes, a partial update is performed on the corresponding row of the regional medical resource availability status matrix, and a timestamp is appended, including: Continuously monitor data change events in the regional medical resource database, and extract the medical institution identifier, changed field type, and changed value carried in the change event when a change event is detected; Based on the medical institution identifier, locate the corresponding row in the regional medical resource availability status matrix, determine the column attribute to be updated based on the changed field type, write the changed value into the column attribute position to be updated in the corresponding row, and append the current time timestamp to that position to complete the partial update of the regional medical resource availability status matrix. Before the regional medical resource availability status matrix is ​​accessed by the scheduling operation, the time interval of each element from the current time is calculated based on the timestamp of each matrix element. For matrix elements whose time interval exceeds the preset validity period, the latest value is retrieved from the regional medical resource database and overwritten. The timestamp of the overwritten element is updated to ensure that each element in the regional medical resource availability status matrix used by the scheduling operation is within the valid time range, and the updated regional medical resource availability status matrix is ​​obtained.

[0040] To ensure that the regional medical resource availability matrix reflects the real-time resource dynamics of each medical institution, an event-driven incremental update mechanism is adopted instead of a full refresh strategy. A continuous monitoring channel for resource data changes is established by deploying database triggers or subscribing to database log streams in the regional medical resource database. When key resource information such as the number of beds, equipment status, and staff attendance of any medical institution in the regional medical resource database changes, the database layer immediately generates a change event and pushes it to the monitoring module.

[0041] Upon receiving a change event, the monitoring module immediately parses the event payload and extracts the core parameters it contains. First, it identifies the medical institution's identifier, which is typically a unique code for the institution, such as a 12-digit institution code. Second, it extracts the change field type, which specifies the particular resource attribute involved in the change, such as "number of ICU beds," "number of available ventilators," or "number of emergency room physicians on duty." Finally, it obtains the changed value, i.e., the latest status value of the updated resource attribute. For example, if the number of ICU beds changes from 18 to 15, the changed value is 15.

[0042] After parameter extraction, the matrix positioning and writing process is initiated. Using the medical institution identifier as the index key, the corresponding row is searched in the regional medical resource availability matrix. Since a row index mapping table was established based on the medical institution identifier during matrix construction, the target row can be quickly located using hash lookup or binary search, keeping the search time complexity constant. After locating the target row, the column attribute to be updated in that row is determined based on the changed field type. For example, if column 5 of the matrix corresponds to the ICU bed count attribute, then column 5 is determined to be the column to be updated when the changed field type is "ICU bed count".

[0043] When performing a value write operation, the modified value is directly overwritten with the original value at the intersection of the corresponding row and column of the matrix. Simultaneously, a timestamp of the current moment is appended to the metadata area associated with that position. The timestamp uses the Unix timestamp format, recording the precise second-level time of the value write. For example, if the write is completed at 10:30:45 on January 15, 2025, the appended timestamp will be 1736917845. This timestamp is stored as metadata bound to the matrix element, forming a "value-timestamp" binary structure, providing a basis for subsequent timeliness verification.

[0044] The aforementioned partial update mechanism performs write operations only on the specific matrix elements that have changed, avoiding a full refresh of the entire matrix. When multiple resource attribute changes occur simultaneously at a medical institution, the monitoring module receives multiple change events. The system processes each event sequentially, updating the corresponding column attribute positions in the matrix, with each updated position independently stamped with a timestamp. This fine-grained incremental update method significantly reduces data synchronization overhead, especially in large-scale regional medical networks containing hundreds of medical institutions, reducing data transmission volume and computational load by over 90% compared to a full refresh.

[0045] Before the scheduling algorithm accesses the regional medical resource availability matrix to perform resource allocation calculations, a timeliness verification and data refresh process is initiated. All elements in the matrix and their associated timestamps are traversed. For each element, the difference between its timestamp and the current time is calculated to obtain the timeliness interval. For example, if an element's timestamp is 1736917845 and the current Unix timestamp is 1736920045, the timeliness interval is 2200 seconds, approximately 36 minutes. The calculated timeliness interval is compared with a preset validity period threshold. The preset validity period is determined based on the dynamic characteristics of the resource type; for relatively stable resources such as the number of beds, it can be set to 60 minutes, while for frequently changing resources such as the number of emergency room physicians on duty, it can be set to 15 minutes.

[0046] When the expiration interval of a matrix element exceeds the preset validity period of its corresponding resource type, it is determined that the element's data may be expired, and the latest value needs to be retrieved from the regional medical resource database. The system constructs a database query statement, specifying the medical institution identifier and resource attribute fields corresponding to the element, and initiates a real-time query request to the regional medical resource database. After the database returns the query results, the latest value is overwritten, replacing the original value at that position in the matrix, and the timestamp at that position is updated to the current time, indicating that the data has been refreshed.

[0047] After refreshing all expired elements in the matrix, all elements in the regional medical resource availability status matrix are ensured to be within the valid timeframe. At this point, the freshness of the matrix data is guaranteed, and the resource status information of each medical institution remains highly consistent with the actual situation. The matrix that has undergone timeliness verification and refreshing is used as the updated regional medical resource availability status matrix and provided to the scheduling algorithm. When the scheduling algorithm performs resource matching and allocation operations based on this matrix, it can ensure that the data used for decision-making is accurate and valid, avoiding the dispatch of patients to medical institutions with no available resources due to data lag, thereby improving the accuracy and reliability of cross-regional medical resource scheduling.

[0048] In practical applications, for example, a regional healthcare network might contain 80 medical institutions, and the available status matrix of regional medical resources might have an 80-row by 15-column structure. If, at a certain moment, five medical institutions simultaneously experience resource changes, an incremental update mechanism only needs to update the corresponding five rows (8 elements) of the matrix, instead of refreshing all 1200 elements. When the scheduling algorithm is about to execute resource allocation calculations, the system checks and finds that the validity interval of 12 elements in the matrix has exceeded the preset expiration date. It only retrieves the latest values ​​from the database for these 12 elements to complete the update, ensuring the overall validity of the matrix data while minimizing data synchronization costs, thus supporting real-time scheduling requirements in high-concurrency scenarios.

[0049] For example, based on the dynamic severity rating and the regional medical resource availability matrix, and constrained by patient transfer time, treatment capacity matching degree, and resource load balance, a cross-institutional collaborative matching and scheduling operation is performed on trauma patients and medical resources to generate an optimal cross-institutional scheduling instruction, including: Based on the dynamic severity rating, the urgency level of treatment and the type of specialist treatment required for each trauma patient are determined. Based on the available bed capacity and specialist treatment suitability of each medical institution in the regional medical resource availability status matrix, a set of candidate matching pairs is generated. For each matching pair in the candidate matching pair set, the estimated transfer time from the patient to the target medical institution is calculated by combining the geographical coordinates in the regional medical resource availability status matrix. The transfer time constraint is verified based on the estimated transfer time and the urgency level of the patient's treatment. Matching pairs that do not meet the transfer time constraint are eliminated, and candidate matching pairs that meet the transfer time constraint are obtained. Among the candidate matching pairs that meet the transfer time constraint, matching pairs whose specialty treatment suitability of the target medical institution does not meet the type of specialty treatment required by the patient are screened out according to the treatment capacity matching degree constraint. In the remaining candidate matching pairs, the balance of the number of patients carried by each medical institution is checked according to the resource load balance degree constraint. Matching combinations whose load difference exceeds the balance constraint are adjusted to generate the optimal cross-institution scheduling instruction that meets all constraints.

[0050] After obtaining the dynamic severity rating and the regional medical resource availability matrix, a cross-institutional collaborative matching and scheduling process is initiated. First, the rating value for each trauma patient in the dynamic severity rating is analyzed and mapped to a treatment urgency classification system. This system categorizes patients into four levels: extremely urgent, highly urgent, moderately urgent, and low urgent. For example, patients with a rating value of 8.5 or higher are classified as extremely urgent, requiring transfer and initiation of treatment within 15 minutes; patients with rating values ​​between 6.5 and 8.5 are classified as highly urgent, with a transfer time limit of 30 minutes; patients with rating values ​​between 4.0 and 6.5 are classified as moderately urgent, with a transfer time limit extended to 60 minutes; and patients with rating values ​​below 4.0 are classified as low urgent, allowing for transfer within 90 minutes.

[0051] The system simultaneously analyzes the injury type characteristics associated with the dynamic severity rating to determine the required specialty treatment type for each patient. Based on the combination of injury location and mechanism, the required specialty treatment type is categorized as follows: neurosurgery for patients with traumatic brain injury and a Glasgow Coma Scale score below 12; thoracic surgery for patients with blunt chest trauma and more than three rib fractures or hemopneumothorax; orthopedic and vascular surgery for patients with pelvic fracture and hemodynamic instability; and general surgery for patients with closed abdominal injury and ultrasound findings of ascites. For patients with multiple injuries, a multi-specialty treatment type combination identifier is generated.

[0052] The process iterates through all medical institution records in the regional medical resource availability matrix, extracting the available bed capacity and specialty treatment suitability identifier for each institution. The specialty treatment suitability is represented by a binary code, indicating the institution's specialty treatment capabilities. For example, the code 10110101 indicates that the institution has the capabilities for neurosurgery, orthopedics, general surgery, and emergency medicine. Bitwise operations are performed to match the required specialty treatment type for each patient with the specialty treatment suitability of each medical institution. If the bitwise AND operation between the institution's specialty code and the patient's requirement code results in all the required specialty bits, the institution is deemed to be able to provide the necessary treatment. Simultaneously, the available bed capacity of the institution is checked to ensure it is greater than 0. Only when the specialty matches and the bed is available is the patient and institution paired as a candidate match. This process iterates through all patient-institution combinations to generate a set of candidate match pairs.

[0053] For each matching pair in the candidate matching pair set, the patient's current geographic coordinates and the target medical institution's geographic coordinates are extracted. The spherical distance is calculated based on the latitude and longitude coordinates of the two points, and the actual surface distance is obtained using the Haversine formula. A road curvature coefficient is applied to the straight-line distance, taking into account the real-time traffic conditions of the regional road network. This coefficient typically ranges from 1.3 to 1.5 in urban areas and from 1.1 to 1.2 in suburban areas. The estimated transfer time is obtained by dividing the corrected actual path distance by the average ambulance speed. The ambulance speed is calculated at 40 km / h in urban areas, 80 km / h on highways, and reduced to 20 km / h in congested areas.

[0054] The calculated estimated transport time is compared with the transport time limit corresponding to the patient's treatment urgency level. For extremely urgent patients, if the estimated transport time exceeds 15 minutes, the match is deemed not to meet the transport time constraint; for highly urgent patients, the match is removed if the estimated transport time exceeds 30 minutes; for moderate and low-urgent patients, the transport time limit thresholds are 60 minutes and 90 minutes, respectively. All candidate matches are iterated through to perform transport time constraint verification, removing matches whose estimated transport time exceeds the corresponding time limit from the candidate set, and retaining candidate matches that meet the transport time constraint to proceed to the next stage of screening.

[0055] Initiate a treatment capacity matching constraint check among candidate matching pairs that meet the transport time constraint. Extract detailed attributes of the specialty treatment suitability of the target medical institution in each matching pair, including the number of physicians in each specialty, the completeness of specialty equipment, and the recent specialty surgery success rate. Quantitatively compare these attributes with the specific requirements of the type of specialty treatment required by the patient. For example, for a severe traumatic brain injury patient requiring neurosurgical treatment, the target institution must have at least 3 neurosurgeons with intracranial pressure monitoring and craniotomy equipment, and a neurosurgical success rate of over 92% in the past 30 days. If any key specialty attribute of the target institution is lower than the patient's demand threshold, the treatment capacity matching is deemed insufficient, and the matching pair is removed from the candidate set.

[0056] Perform resource load balancing constraint verification on the remaining candidate matching pairs. Count the number of patients assigned to each medical institution in the current candidate matching scheme, and calculate the occupancy rate of each institution's patient capacity relative to its total bed capacity. Define the load balancing index as the standard deviation of the occupancy rates of all institutions; when the standard deviation exceeds 0.25, the load distribution is considered unbalanced. Identify the institutions with the highest and lowest occupancy rates. From the patients assigned to the high-occupancy institutions, select patients with relatively low urgency levels whose transfer time to the low-occupancy institutions still meets the constraints, and adjust their matching relationships to the low-occupancy institutions. Iteratively execute the adjustment operation until the load balancing index drops below 0.25 or there is no room for further adjustment.

[0057] After completing all constraint verification and adjustment, the final matched medical institution for each trauma patient is determined. An optimal cross-institutional dispatch instruction set is generated, with each instruction containing a unique patient identifier, the name of the target medical institution, the estimated transfer time, the required specialty treatment type, and the coordinates of key nodes along the suggested transfer route. The dispatch instruction set is distributed to the emergency dispatch center and relevant medical institutions, triggering ambulance dispatch and reception preparation processes, achieving efficient cross-institutional collaborative treatment dispatch for trauma patients.

[0058] For example, based on resource load balance constraints, a balance check is performed on the number of patients carried by each medical institution, and matching combinations where the load difference exceeds the balance constraint are adjusted to generate an optimal cross-institution scheduling instruction that satisfies all constraints, including: The total number of patients assigned to each target medical institution in the remaining candidate matching pairs is counted. The total number of patients is divided by the available bed capacity of the corresponding medical institution to obtain the current load occupancy rate of each medical institution. All medical institutions in the region are then sorted according to their load occupancy rates. The difference between the maximum and minimum load occupancy rates of each medical institution in the region is calculated to obtain the load difference degree. The load difference degree is compared with the resource load balance constraint. For the matching combination where the load difference degree exceeds the balance constraint, the medical institution with the highest load occupancy rate and the transferable patients it carries are identified. The transferable patients are reassigned from the medical institution with the highest load difference to other medical institutions that meet the constraints of transfer time and treatment capacity matching degree and have the lowest current load occupancy rate. The load difference of each medical institution is recalculated for the reassigned matching combination. The above reassignment operation is repeated until the load difference of all medical institutions meets the balance constraint. All the finally determined matching pairs are encapsulated into the optimal cross-institution scheduling instruction.

[0059] After generating initial candidate matching pairs, it is necessary to perform a patient load balance check on each medical institution to avoid situations where some institutions are overloaded while others are idle. The actual load on each institution can be quantified by counting the total number of patients assigned to each target medical institution in the current matching scheme. Specifically, all remaining candidate matching pairs are iterated, and the number of patients assigned to each target medical institution is accumulated. For example, in a region containing 5 medical institutions, after initial matching, Hospital A is assigned 12 patients, Hospital B is assigned 8 patients, Hospital C is assigned 15 patients, Hospital D is assigned 6 patients, and Hospital E is assigned 9 patients.

[0060] To more accurately reflect the workload of each medical institution, it is necessary to compare and calculate the total number of patients with the available bed capacity of the corresponding institution. Available bed capacity data for each medical institution is obtained from the medical resource management system; this data should be updated in real time to reflect the actual number of beds available to receive patients. The total number of patients allocated to each medical institution is divided by its available bed capacity to obtain the current load factor of each institution. The calculated load factor is then used as the ranking criterion to sort all medical institutions in the region from highest to lowest.

[0061] To quantify the degree of balance in the allocation of medical resources within a region, the difference between the maximum and minimum load occupancy rates of each medical institution was calculated. The greater the load difference, the more uneven the allocation of medical resources.

[0062] The calculated load difference is compared with the preset resource load balance constraint. The resource load balance constraint is usually set according to the regional medical resource management policy, such as requiring that the load difference between medical institutions in the region should not exceed 30%. If the calculated load difference of 53.6% exceeds the 30% balance constraint, it indicates that there is an imbalance in load distribution in the current matching combination, and adjustment operations need to be performed.

[0063] For matched combinations where the load difference exceeds the equilibrium constraint, the medical institution with the highest load occupancy rate is identified. For example, Hospital C has the highest load occupancy rate at 75%, making it a key target for load reduction. Further analysis of all patients currently at this medical institution identifies transferable patients. Identification of transferable patients requires comprehensive consideration of the patient's urgency, transfer feasibility, and impact on scheduling effectiveness. Generally, patients with relatively low urgency levels, relatively stable conditions, and low transfer risks are prioritized for transfer. For example, among the 15 patients at Hospital C, 3 patients with moderate urgency levels and stable conditions can be identified as transferable patients.

[0064] A re-matching process is performed on the identified transferable patients. Other medical institutions within the region are traversed to select candidate receiving institutions that meet both the transfer time constraint and the treatment capacity matching constraint. The transfer time constraint ensures that the patient can arrive at the new medical institution within the specified time, while the treatment capacity matching constraint ensures that the receiving institution has the professional capability to handle the patient's condition. Among the candidate receiving institutions that meet both constraints, the institution with the lowest current load factor is prioritized for re-matching. In the example, Hospital D has the lowest load factor of 21.4% and has been verified to meet the transfer time and treatment capacity requirements. Therefore, two transferable patients from Hospital C are reassigned to Hospital D, and another transferable patient is reassigned to Hospital B.

[0065] After the reallocation operation, the load occupancy rate and load variance of each medical institution were recalculated. After the adjustment, Hospital C's patient count decreased to 13, and its load occupancy rate became 13 divided by 20, which equals 0.65, or 65%. Hospital D's patient count increased to 8, and its load occupancy rate became 8 divided by 28, which equals approximately 0.286, or 28.6%. Hospital B's patient count increased to 9, and its load occupancy rate became 9 divided by 25, which equals 0.36, or 36%. The recalculated maximum load occupancy rate was 65%, the minimum load occupancy rate was 28.6%, and the load variance was 36.4%, still exceeding the 30% equilibrium constraint.

[0066] Therefore, the above reallocation operation needs to be repeated. The medical institution with the highest current load factor (65%) is identified again as Hospital C. One patient who can be transferred is identified from its patient load and reallocated to Hospital D, which has a lower load factor. After the adjustment, the number of patients in Hospital C decreases to 12, with a load factor of 60%; the number of patients in Hospital D increases to 9, with a load factor of approximately 32.1%. At this point, the maximum load factor is 60%, the minimum load factor is 32%, and the load difference is 28%, satisfying the 30% equilibrium constraint.

[0067] Once the load disparity across all healthcare institutions meets the equilibrium constraint, the reallocation operation terminates. All finalized matching pairs are then encapsulated to generate an optimal cross-institutional scheduling instruction. This instruction is stored in a structured data format, containing complete information for each patient, including their identification information, current location, target healthcare institution, estimated transfer time, priority level, and any special medical needs.

[0068] The optimal cross-agency scheduling instruction is distributed to the corresponding medical node via the cloud computing platform. The corresponding medical node then executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction, including: The cloud computing platform, based on the medical node identifier in the optimal cross-institutional scheduling instruction, splits the optimal cross-institutional scheduling instruction into sub-instructions corresponding to each target medical node, sends each sub-instruction to the corresponding medical node, confirms the delivery status of each sub-instruction, and retransmits sub-instructions that do not return a confirmation response within a preset time limit. After receiving the sub-instruction, the corresponding medical node parses the patient information, estimated arrival time and required treatment resource type in the sub-instruction, issues a resource preparation instruction to the corresponding department based on the estimated arrival time, completes the preparatory deployment of the corresponding treatment resources to the designated reception area, and reports the deployment status of the treatment resources to the cloud computing platform. After aggregating the deployment status of medical resources reported by each medical node, the cloud computing platform identifies medical nodes with resource gaps and sends resource replenishment requests to other medical nodes in the region with available resources, thereby completing the cross-node collaborative allocation of trauma treatment resources in the region.

[0069] In the actual operation of the trauma emergency dispatch system, after receiving the optimal cross-institutional dispatch instruction, the cloud computing platform immediately initiates the instruction distribution mechanism. The cloud computing platform first performs structured parsing of the optimal cross-institutional dispatch instruction, extracting all medical node identifiers contained within. These identifiers adopt a nationally unified medical institution coding system; for example, the code for a tertiary hospital is "310100-A-001," and the code for a secondary hospital is "310100-B-023." Based on the number of medical nodes involved in the instruction, the cloud computing platform breaks down the original dispatch instruction according to the node dimension, forming independent sub-instructions for each target medical node. Each sub-instruction contains the specific tasks that the node needs to perform, including the injury code of the patient to be received, the estimated arrival time window, the type of department to be prepared, and a specific list of treatment resources.

[0070] The broken-down sub-instructions are distributed via the cloud computing platform's push service, employing an asynchronous communication mechanism based on message queues. Each sub-instruction is assigned a unique message identifier upon transmission and has a set expected confirmation timeout, typically 15 to 30 seconds. The cloud computing platform establishes a delivery status monitoring record for each sub-instruction, tracking the response status of each medical node in real time. When a medical node receives a sub-instruction, its locally deployed trauma management terminal immediately returns a confirmation response to the cloud computing platform, carrying the message identifier, reception timestamp, and the node's current status code in the response message. If the cloud computing platform does not receive a confirmation response for a sub-instruction within the preset timeout, the system determines that the instruction may have failed to be delivered due to network fluctuations or terminal malfunction, and immediately initiates a retransmission mechanism, retransmitting the instruction through a backup communication link or SMS gateway to ensure that critical scheduling instructions are not missed.

[0071] Upon successfully receiving the sub-instruction, the trauma management terminal at the target medical node initiates a local parsing process. The terminal system extracts basic patient information fields from the sub-instruction message, including gender, age, and injury severity code. For example, an injury severity rating of "ISS-25" indicates a severity score of 25, classifying it as severe trauma. The system simultaneously acquires estimated arrival time data, presented as a time range, such as "arriving in 18 to 22 minutes," providing a time margin for resource preparation. The required treatment resource types are given in a standardized coded list format; for example, "RES-OR-001" indicates the need for one operating room, "RES-BLD-AB-400" indicates the need for 400 ml of AB type blood, and "RES-DOC-ORTHO-2" indicates the need for two orthopedic surgeons on standby.

[0072] The medical node system calculates the resource preparation activation time based on the parsed arrival time. If the patient is expected to arrive in 20 minutes, considering that operating room preparation takes 8 to 10 minutes, medical staff assembly takes 5 minutes, and instrument sterilization takes 6 minutes, the system will immediately issue resource preparation instructions to the relevant departments upon receiving the instruction. These preparation instructions are distributed to the operating room dispatch center, blood bank management system, doctor on-call dispatch system, and nursing scheduling system via the hospital's internal LAN or wireless communication system. Upon receiving the preparation instructions, the operating room dispatch center immediately arranges for the designated operating room to perform preoperative preparations, including adjusting the indoor temperature and humidity to a suitable range, turning on the shadowless lamps and monitoring equipment, and laying out sterile operating table drapes. The blood bank management system extracts the corresponding blood products from the cold storage according to the blood type and usage requirements, completes cross-matching tests, and transfers the blood bags to the operating room's backup area.

[0073] The on-call orthopedic surgeon receives a standby notification via mobile terminal, immediately suspends non-urgent work, and proceeds to the designated reception area. Nursing staff simultaneously receive instructions, carrying trauma kits, intravenous access equipment, and vital sign monitoring devices to the emergency trauma resuscitation unit. The medical node's supplies management system automatically retrieves the required equipment list, confirms the inventory status and location of orthopedic surgical instrument kits, external fixators, ventilators, and other equipment, and notifies the equipment department staff to transport the equipment to the designated area. Throughout the resource preparation process, each subsystem reports its execution status to the trauma treatment management terminal after completing its respective task. The terminal system then compiles these reports into a resource deployment completion report.

[0074] The report includes the actual deployment status of various resources. For example, operating room 2 is ready and ready for use; 400 ml of AB-type red blood cell suspension has been delivered to the operating room; two orthopedic attending physicians are in place; one anesthesiologist is in place; three nursing staff are in place; and the C-arm X-ray machine is in place and has completed self-checking. If a shortage of any resource is found during the resource preparation process, such as insufficient AB-type blood inventory in the blood bank (only 280 ml available, while the requirement is 400 ml), the system will indicate the type and quantity of the resource shortage in the deployment status report. The medical node will report the complete deployment status report of the treatment resources to the cloud computing platform through an encrypted channel. The reported data is in a structured format to facilitate automated processing by the cloud platform.

[0075] After receiving deployment status reports from various medical nodes within the region, the cloud computing platform initiates a resource gap analysis module. The platform categorizes all reported data by medical node, checking the resource availability indicators in each node's report. When a medical node is identified as having a resource gap, the platform extracts the type, quantity, and urgency level of the missing resource. For example, if a medical node is identified as having a shortage of 120 ml of type AB blood, the platform immediately queries the blood inventory data of other medical nodes within the region. The platform then filters from the real-time synchronized regional blood inventory database to identify medical nodes holding type AB blood with remaining quantities exceeding the safety stock threshold, calculating the geographical distance and estimated transport time between each candidate node and the node in need.

[0076] The cloud computing platform comprehensively considers the transfer time, the remaining blood volume, and the current load of the supply nodes to select the optimal resource replenishment node. The platform sends a resource replenishment request to this node, clearly indicating in the request message that the type of resource to be allocated is "AB type red blood cell suspension", the allocation volume is "150 ml" (with a safety margin), the receiving node identifier, the transfer time limit, and other information.

[0077] A second aspect of the present invention provides a cloud-based regional major trauma data processing system, comprising: The data acquisition unit is used to acquire the raw data of trauma patients reported by each medical node in the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform, generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. The resource status unit is used to extract the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution from the regional medical resource database of the cloud computing platform based on the dynamic severity rating, and to construct a regional medical resource availability status matrix. The collaborative matching unit is used to perform cross-institutional collaborative matching scheduling operations on trauma patients and medical resources based on the dynamic severity rating and the regional medical resource availability status matrix, with patient transfer time, treatment capacity matching degree and resource load balance as constraints, and generate the optimal cross-institutional scheduling instruction. The instruction issuing unit is used to issue the optimal cross-agency scheduling instruction to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

[0078] 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.

[0079] 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.

[0080] 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.

[0081] 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. A cloud computing-based method for processing regional major trauma data, characterized in that, include: The system acquires raw data of trauma patients reported by medical nodes within the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform. It generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. Based on the dynamic severity rating, the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix. Based on the dynamic severity rating and the regional medical resource availability matrix, and with patient transfer time, treatment capacity matching degree, and resource load balance as constraints, cross-institutional collaborative matching and scheduling operations are performed on trauma patients and medical resources to generate optimal cross-institutional scheduling instructions. The optimal cross-agency scheduling instruction is sent to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

2. The method according to claim 1, characterized in that, Parallel distributed scoring calculations are performed on a cloud computing platform to generate an initial score for each patient based on trauma severity scoring rules. The weights of each dimension are then adaptively updated according to real-time changes in the initial score to obtain a dynamic severity rating for each trauma patient, including: The multidimensional physiological parameters are divided into multiple scoring sub-tasks according to the dimension category and distributed to multiple parallel computing nodes of the cloud computing platform. Each parallel computing node performs independent scoring calculations on the assigned scoring sub-tasks according to the trauma severity scoring rules to obtain the scores for each dimension. At the aggregation node, a weighted aggregation of the scores for each dimension is performed based on the current weight of each dimension to obtain the initial score for each patient; Based on the change in the initial score within a continuous evaluation period, the contribution of each scoring dimension to the change in the initial score is identified. For dimensions whose contribution exceeds the current weight representation range, the corresponding weight is increased; for dimensions whose contribution is less than the current weight representation range, the corresponding weight is decreased, thus completing the adaptive update of the scoring weights of each dimension. The scores of each dimension are re-weighted and aggregated according to the adaptively updated score weights to obtain an updated score. The updated score is then mapped to the corresponding severity level according to the severity grading mapping rule to obtain the dynamic severity rating for each trauma patient.

3. The method according to claim 1, characterized in that, The real-time bed occupancy rate, specialty treatment capacity index, and geographical coordinates of each medical institution are extracted from the regional medical resource database of the cloud computing platform to construct a regional medical resource availability status matrix, including: The current bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution are obtained by polling the regional medical resource database according to the medical institution identifier. The available bed capacity of each medical institution is calculated based on the current bed occupancy rate. The degree of suitability of each medical institution for the treatment of various types of trauma is determined based on the specialty treatment capacity index. Using the medical institution identifier as the row index and the available bed capacity, the degree of adaptation of each specialty treatment, and the geographical coordinates as column attributes, the acquired medical institution resource data are written into the corresponding positions in the matrix to obtain the regional medical resource availability status matrix. When resource data in the regional medical resource database changes, a partial update is performed on the available bed capacity and specialty treatment suitability of the corresponding row in the regional medical resource availability status matrix, and a timestamp is added to the updated matrix elements to ensure that the regional medical resource availability status matrix is ​​consistent with the data status of the regional medical resource database, thus obtaining the updated regional medical resource availability status matrix.

4. The method according to claim 3, characterized in that, When resource data in the regional medical resource database changes, a partial update is performed on the corresponding row of the regional medical resource availability status matrix, and a timestamp is appended, including: Continuously monitor data change events in the regional medical resource database, and extract the medical institution identifier, changed field type, and changed value carried in the change event when a change event is detected; Based on the medical institution identifier, locate the corresponding row in the regional medical resource availability status matrix, determine the column attribute to be updated based on the changed field type, write the changed value into the column attribute position to be updated in the corresponding row, and append the current time timestamp to that position to complete the partial update of the regional medical resource availability status matrix. Before the regional medical resource availability status matrix is ​​accessed by the scheduling operation, the time interval of each element from the current time is calculated based on the timestamp of each matrix element. For matrix elements whose time interval exceeds the preset validity period, the latest value is retrieved from the regional medical resource database and overwritten. The timestamp of the overwritten element is updated to ensure that each element in the regional medical resource availability status matrix used by the scheduling operation is within the valid time range, and the updated regional medical resource availability status matrix is ​​obtained.

5. The method according to claim 1, characterized in that, Based on the dynamic severity rating and the regional medical resource availability matrix, and constrained by patient transfer time, treatment capacity matching degree, and resource load balance, a cross-institutional collaborative matching and scheduling operation is performed on trauma patients and medical resources to generate optimal cross-institutional scheduling instructions, including: Based on the dynamic severity rating, the urgency level of treatment and the type of specialist treatment required for each trauma patient are determined. Based on the available bed capacity and specialist treatment suitability of each medical institution in the regional medical resource availability status matrix, a set of candidate matching pairs is generated. For each matching pair in the candidate matching pair set, the estimated transfer time from the patient to the target medical institution is calculated by combining the geographical coordinates in the regional medical resource availability status matrix. The transfer time constraint is verified based on the estimated transfer time and the urgency level of the patient's treatment. Matching pairs that do not meet the transfer time constraint are eliminated, and candidate matching pairs that meet the transfer time constraint are obtained. Among the candidate matching pairs that meet the transfer time constraint, matching pairs whose specialty treatment suitability of the target medical institution does not meet the type of specialty treatment required by the patient are screened out according to the treatment capacity matching degree constraint. In the remaining candidate matching pairs, the balance of the number of patients carried by each medical institution is checked according to the resource load balance degree constraint. Matching combinations whose load difference exceeds the balance constraint are adjusted to generate the optimal cross-institution scheduling instruction that meets all constraints.

6. The method according to claim 5, characterized in that, Based on resource load balance constraints, a balance check is performed on the number of patients carried by each medical institution. Matching combinations where load differences exceed the balance constraints are adjusted, and an optimal cross-institution scheduling instruction that satisfies all constraints is generated, including: The total number of patients assigned to each target medical institution in the remaining candidate matching pairs is counted. The total number of patients is divided by the available bed capacity of the corresponding medical institution to obtain the current load occupancy rate of each medical institution. All medical institutions in the region are then sorted according to their load occupancy rates. The difference between the maximum and minimum load occupancy rates of each medical institution in the region is calculated to obtain the load difference degree. The load difference degree is compared with the resource load balance constraint. For the matching combination where the load difference degree exceeds the balance constraint, the medical institution with the highest load occupancy rate and the transferable patients it carries are identified. The transferable patients are reassigned from the medical institution with the highest load difference to other medical institutions that meet the constraints of transfer time and treatment capacity matching degree and have the lowest current load occupancy rate. The load difference of each medical institution is recalculated for the reassigned matching combination. The above reassignment operation is repeated until the load difference of all medical institutions meets the balance constraint. All the finally determined matching pairs are encapsulated into the optimal cross-institution scheduling instruction.

7. The method according to claim 1, characterized in that, The optimal cross-agency scheduling instruction is distributed to the corresponding medical node via the cloud computing platform. The corresponding medical node then executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction, including: The cloud computing platform, based on the medical node identifier in the optimal cross-institutional scheduling instruction, splits the optimal cross-institutional scheduling instruction into sub-instructions corresponding to each target medical node, sends each sub-instruction to the corresponding medical node, confirms the delivery status of each sub-instruction, and retransmits sub-instructions that do not return a confirmation response within a preset time limit. After receiving the sub-instruction, the corresponding medical node parses the patient information, estimated arrival time and required treatment resource type in the sub-instruction, issues a resource preparation instruction to the corresponding department based on the estimated arrival time, completes the preparatory deployment of the corresponding treatment resources to the designated reception area, and reports the deployment status of the treatment resources to the cloud computing platform. After aggregating the deployment status of medical resources reported by each medical node, the cloud computing platform identifies medical nodes with resource gaps and sends resource replenishment requests to other medical nodes in the region with available resources, thereby completing the cross-node collaborative allocation of trauma treatment resources in the region.

8. A cloud-based regional major trauma data processing system, used to implement the method as described in any one of claims 1-7, characterized in that, include: The data acquisition unit is used to acquire the raw data of trauma patients reported by each medical node in the region. Based on the multidimensional physiological parameters in the raw data of trauma patients, it performs parallel distributed scoring calculation on the cloud computing platform, generates an initial score for each patient according to the trauma severity scoring rules, and performs adaptive updates on the scoring weights of each dimension according to the real-time changes of the initial score to obtain the dynamic severity rating of each trauma patient. The resource status unit is used to extract the real-time bed occupancy rate, specialty treatment capacity index and geographical coordinates of each medical institution from the regional medical resource database of the cloud computing platform based on the dynamic severity rating, and to construct a regional medical resource availability status matrix. The collaborative matching unit is used to perform cross-institutional collaborative matching scheduling operations on trauma patients and medical resources based on the dynamic severity rating and the regional medical resource availability status matrix, with patient transfer time, treatment capacity matching degree and resource load balance as constraints, and generate the optimal cross-institutional scheduling instruction. The instruction issuing unit is used to issue the optimal cross-agency scheduling instruction to the corresponding medical node via the cloud computing platform, and the corresponding medical node executes the preparation for receiving trauma patients and the allocation of treatment resources according to the optimal cross-agency scheduling instruction.

9. 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 7.

10. 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 7.

Images