Call-to-treatment time optimization method and guidance system

Abstract

The invention provides a closed-loop method for optimizing call to treatment time and a guidance system. The system carries out real-time evaluation and optimization decision making on patient conditions, ambulance positions, traffic conditions and medical resources, and dynamic adjustment is carried out in the execution process. And meanwhile, first-aid measures are provided for the site through remote guidance, and feedback information is collected. And after the event is finished, an actual execution result is analyzed, parameter updating and structure improvement are performed on the optimization decision model and the guidance strategy, and data closed-loop feedback is formed. Through continuous iterative optimization, the overall efficiency and quality from help calling to treatment can be remarkably improved, the self-adaptive ability of the system to cope with a dynamic change environment is enhanced, and the patient treatment success rate and the medical resource utilization rate are improved.

Description

Technical Field

[0001] The present invention belongs to the field of medical rescue and emergency response, and specifically relates to a scheduling and guidance technical method and system for optimizing the time from calling for help to treatment, so as to improve the efficiency of emergency resource allocation and the speed of emergency medical response. Background Art

[0002] With the continuous acceleration of urbanization, society has put forward higher requirements for the emergency medical rescue system. In emergency events, the time difference between the caller's request to the emergency center and the arrival of professional medical staff at the scene and effective treatment (i.e., "call-to-treatment time") is crucial. The length of the call-to-treatment time is directly related to the patient's survival rate and treatment effect. However, the current emergency dispatch and guidance often have the following problems: 1. The dispatch process is complicated. The existing emergency dispatch relies on manual experience to dispatch resources, and it is difficult to dynamically optimize the dispatch strategy according to the real-time resource distribution and road traffic conditions in a timely and accurate manner; 2. Information asymmetry. The emergency center has insufficient grasp of the on-site situation. The decision-making process lacks overall consideration of real-time road conditions, ambulance location, availability of medical staff, and expected processing time; 3. Lack of intelligent auxiliary guidance. There is a lack of intelligent system to automatically optimize and assist dispatch decisions, and the real-time guidance of the caller and on-site witnesses in the implementation of professional emergency measures is insufficient, resulting in the precious "golden minutes" before the arrival of the ambulance personnel. Often, it is not reasonably utilized.

[0003] In response to the above problems, it is necessary to propose a technical method and system that can dynamically optimize the time from calling for help to treatment. Based on intelligent analysis and decision support of real-time data, resource scheduling optimization and on-site first aid guidance can be carried out from the time the call is initiated to shorten the overall treatment process to the greatest extent possible. Summary of the invention

[0004] The purpose of the present invention is to provide a call-to-treatment time optimization method and guidance system, which comprehensively considers the distribution of emergency resources, road traffic conditions, patient status assessment and on-site emergency measures guidance, and uses intelligent algorithms and data analysis capabilities to provide real-time decision-making suggestions and assistance to the emergency dispatch center and front-end rescue forces, thereby achieving global optimization of the call-to-treatment time.

[0005] To achieve the above object, the present invention provides a method for optimizing the call-to-help time, comprising the following steps:

[0006] (1) Data acquisition: real-time acquisition includes but is not limited to the geographic location of the ambulance, road traffic conditions, availability of emergency medical personnel, hospital department reception capacity, historical dispatch data, and symptom description information when the patient calls for help;

[0007] (2) Status assessment: make a preliminary estimate of the patient's possible injury type and urgency, and analyze the resource status of available ambulances and paramedics;

[0008] (3) Dispatch optimization: Based on the state evaluation results, the ambulance dispatch plan is calculated using an optimization algorithm (such as integer programming, genetic algorithm, or deep reinforcement learning method) to obtain the optimal or suboptimal task allocation and driving path planning;

[0009] (4) Real-time tracking and dynamic adjustment: After the ambulance is dispatched, the dispatch plan is dynamically adjusted based on real-time traffic changes, patient status updates, and other unexpected factors;

[0010] (5) On-site guidance: Before the arrival of rescue forces, provide basic first aid measures guidance to the caller or on-site personnel through voice, text or video, including cardiopulmonary resuscitation, hemostasis, correct positioning, etc., to stabilize the patient's condition;

[0011] (6) Data storage and analysis: The data, decision-making plans and processing results in the above process are recorded and archived for later analysis and model optimization.

[0012] The call-to-treatment time optimization guidance system based on the above method includes:

[0013] Data collection and fusion module: used to collect and integrate ambulance location data, traffic data, hospital admission data, historical dispatch data, and patient condition description information.

[0014] Status Assessment Module: Estimates the possible severity of the patient's condition and analyzes and evaluates the available ambulance and medical staff resources.

[0015] Optimization decision module: Based on the preset optimization goals and constraints, the optimization algorithm is used to generate resource dispatch and route planning solutions.

[0016] Dynamic adjustment module: tracks real-time changing information and modifies established scheduling plans.

[0017] Remote guidance module: provides real-time guidance on emergency rescue measures on site, including voice calls, video connections, image displays and text prompts.

[0018] Storage and recording module: stores the entire process data, decision-making strategies and treatment results for subsequent statistical analysis and model iteration optimization.

[0019] Through the above method and system, the overall time from receiving a call for help to implementing treatment can be significantly shortened, the utilization efficiency and rationality of allocation of emergency resources can be improved, and emergency guidance can be provided to patients before arrival, thereby maximizing the success rate of emergency treatment and the quality of treatment.

[0020] The beneficial effects of the present invention include:

[0021] Intelligent optimization: By introducing intelligent algorithms to optimize ambulance dispatch and route planning, the idle rate and dispatch delays can be reduced.

[0022] Global coordination: Make global decisions based on real-time data to improve the rationality of resource allocation.

[0023] Reduce delays: Be sensitive to real-time changes and dynamically adjust scheduling plans to reduce delays caused by traffic congestion and resource constraints.

[0024] Timely guidance: Provide professional guidance before the ambulance arrives to improve the effectiveness of the patient's self-rescue or rescue measures.

[0025] Furthermore, real-time AI voice analysis: When the 120 emergency center receives a call for help, NLP (natural language processing) + AI voice analysis automatically identifies the patient's condition.

[0026] Automatic classification of urgency: Based on keyword extraction (such as "chest pain" and "dyspnea") and combined with historical case data, the emergency level is intelligently classified (red / yellow / green).

[0027] Reduce information entry time, optimize traditional manual inquiry process, and enable the emergency center to complete condition classification within 5 seconds.

[0028] Technical Effect: Shorten call response time: reduce condition identification and classification from 60 seconds to 10 seconds. Improve dispatch accuracy: reduce the number of low-priority patients occupying emergency resources and improve the utilization rate of emergency resources.

[0029] Furthermore, ambulance dispatch has been upgraded from manual dispatch to AI intelligent matching, based on: the nearest ambulance; the ambulance with the highest medical capability match (e.g., emergency vehicles with ECMO equipment are suitable for myocardial infarction patients); and the professional matching of medical staff in the vehicle (e.g., trauma rescue vs. cardiovascular emergency). The dispatch algorithm is optimized based on big data to ensure the dispatch of the best rescue resources.

[0030] Technical effect: Average dispatch time is reduced by 30-50 seconds. Ambulance matching accuracy is increased by 30%, reducing incorrect resource allocation.

[0031] Furthermore, ambulance navigation is not only based on GPS+real-time traffic data, but also combines the following data sources for multi-dimensional route optimization: Traffic flow prediction: Use AI algorithms to analyze road congestion trends and avoid possible traffic jams. Dynamic traffic light control: Connect with the smart traffic signal system to provide a green channel for ambulances (such as automatically extending the green light time). Weather impact analysis: Combine meteorological data to optimize routes and avoid extreme weather impacts.

[0032] Technical effect: The time to arrive at the scene is shortened by 20-30%, ensuring that the ambulance reaches the patient's location faster.

[0033] Furthermore, 5G telemedicine supports: Through 5G+ high-definition video, the camera in the ambulance transmits the patient's vital signs data (such as electrocardiogram, blood oxygen, and blood pressure) in real time. Hospital emergency department experts can view the data in real time and remotely guide rescue personnel to perform pre-hospital emergency treatment (such as whether to use electric defibrillation and the best airway management strategy).

[0034] AI automatic assessment system: The ambulance is equipped with an AI condition assessment module, which automatically generates emergency plan suggestions based on the patient's vital signs + medical record data. Example: When the risk of myocardial infarction is detected, AI automatically instructs emergency personnel to take oral aspirin and notifies the hospital through the remote system to prepare for PCI surgery.

[0035] Technical effects:

[0036] The time for emergency treatment is reduced by 30-60 seconds, and some conditions can be treated in the ambulance first. Hospitals are well prepared in advance to reduce the time for rescue after arrival.

[0037] Furthermore, the AI ​​system automatically sends patient information to the nearest optimal hospital: Early triage: Before the patient arrives, the hospital emergency room can obtain vital signs and preliminary diagnosis information, and arrange emergency doctors and wards in advance. Hospital load balancing: If the first-choice hospital is full, the system automatically recommends a second-priority hospital with a lower load to ensure that patients are not delayed in treatment due to crowded hospitals.

[0038] Technical effect: Reduce the waiting time of patients after arriving at the hospital (shortened by 5-10 minutes on average). Increase the efficiency of hospital reception by 20%-30%. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flowchart of a process for applying for an embodiment of the present invention; DETAILED DESCRIPTION

[0040] The following is further described in detail through specific implementation methods:

[0041] In a typical embodiment of the present invention, the system can be deployed in a city-level emergency command center and realize data interconnection and interaction with several emergency hospitals, vehicle dispatch centers, traffic data platforms and on-site terminal devices. The system hardware and software can be composed of the following modules and terminals:

[0042] Data acquisition and fusion module

[0043] Hardware configuration:

[0044] Data servers (such as Dell PowerEdge R740xd or Inspur NF5280M5 rack servers) are equipped with high-performance CPUs (Intel Xeon series), large-capacity memory (more than 64GB) and high-speed SSD storage.

[0045] Data exchange gateways (Cisco Catalyst series switches or Huawei CloudEngine series switches) are used to connect various data sources.

[0046] Data source and agreement:

[0047] The ambulance location data is transmitted back to the central server by the GPS positioning terminal installed on the ambulance (such as Beidou / GPS dual-mode positioning terminal, brand such as UniStrong UG902) through the 4G / 5G cellular network (supporting LTE Cat 6 and above standards) or private network communication protocol (such as PDT / DMR dedicated digital trunking communication protocol).

[0048] Traffic status data can be obtained from a traffic data platform (such as the Urban Traffic Big Data Center, using Http RESTful API) and transmitted encrypted via the HTTPS protocol.

[0049] Medical resource availability data (hospital bed status, emergency department load, and doctor on-duty information) is connected to the command center's data server through the hospital's internal HIS system. The HL7 (Health Level Seven) international standard protocol or FHIR (Fast Healthcare Interoperability Resources) standard can be used to achieve secure data interaction through medical data middleware (such as InterSystems HealthShare).

[0050] The patient's call for help is accessed by the call center (120 emergency telephone system), connected to the CTI (Computer Telephony Integration) system through SIP (Session Initiation Protocol), and the structured information (such as address, preliminary symptom description) is transmitted to the data server via HTTPS or MQTT.

[0051] Data fusion processing:

[0052] The data server runs data fusion and cleaning software (Apache Flink or Spark Streaming can be used) to integrate data from different sources and formats in real time, and store it in an integrated database (such as PostgreSQL or MySQL distributed cluster) to provide standardized data input for subsequent status evaluation and optimization decisions.

[0053] Condition Assessment Module

[0054] Hardware configuration:

[0055] The status assessment module can be run on the same data center server, or deployed on a dedicated computing node (such as Dell PowerEdge R640).

[0056] Software environment and model:

[0057] The status assessment module runs on a Python-based service framework (such as FastAPI or Flask) and is quickly deployed and expanded through Docker containers.

[0058] The model can be implemented using machine learning, rule engines, and knowledge base technologies. For example, a neural network model trained with TensorFlow or PyTorch can be used to estimate the severity of a patient's symptoms.

[0059] Evaluation Process:

[0060] After receiving the real-time data provided by the data fusion module, the status assessment module conducts a comprehensive analysis based on the pre-trained disease classification model (such as using a multi-layer perceptron model to predict the urgency of symptom text) and resource status rules (such as ambulance dispatch frequency, hospital admission capacity threshold, etc.), and outputs the patient's urgency level, priority allocation recommendations and preliminary resource allocation strategies.

[0061] Optimizing decision-making modules

[0062] Hardware configuration:

[0063] The optimization decision module can be deployed on a server with strong computing power (such as Dell PowerEdge R840 equipped with GPU or multi-core CPU) to handle complex optimization problem calculations.

[0064] Algorithm and software environment:

[0065] The optimization decision module uses mathematical optimization solvers (such as Gurobi, CPLEX) or meta-heuristic algorithm frameworks (such as Google OR-Tools), and can be combined with deep reinforcement learning algorithms (TensorFlow / PyTorch implementation) for dynamic decision-making.

[0066] The communication protocol uses gRPC or REST API to interact with the status assessment module, obtain input parameters (patient status, resource status) and return the initial ambulance dispatch plan and driving route planning.

[0067] Decision content:

[0068] This module optimally matches and solves the routes of ambulances based on distance, road congestion, vehicle resource distribution and hospital reception capacity to ensure that they arrive at the scene in the shortest time.

[0069] Dynamic adjustment module

[0070] Hardware configuration:

[0071] It can share server nodes with the optimization decision module or be deployed independently.

[0072] Software implementation and protocol:

[0073] The dynamic adjustment module periodically calls the optimization algorithm to perform local optimization adjustments based on real-time data streams (such as the current location of the vehicle and road emergency information). It receives real-time update events through lightweight message queues (such as Apache Kafka or RabbitMQ) and feeds back the latest solutions to the decision-making front end in the form of WebSocket or gRPC bidirectional streaming.

[0074] Functional description:

[0075] The dynamic adjustment module continuously monitors the execution of decisions. Once a deviation is detected (such as a detour of the convoy, a temporary roadblock, or a worsening patient condition), it triggers rapid replanning and provides the executors and the site with the latest routes or plans for additional rescue forces.

[0076] Remote guidance module

[0077] Hardware configuration:

[0078] The front end of the remote guidance module can be a multimedia workstation in the emergency command center (such as equipped with a Huawei TE10 video terminal, a Logitech C930e camera, and a professional microphone), or a mobile tablet terminal for medical staff (such as an iPad Pro 2020, supporting 4G / 5G / Wi-Fi).

[0079] Communication protocol and terminal equipment:

[0080] Communication with the scene can be achieved through VoLTE calls, WebRTC video calls or RTP / RTSP video streaming transmission. The on-site terminal can be the rescue witness's smartphone (such as Huawei Mate series or Apple iPhone), installed with a dedicated app or accessed through an H5 page. Encryption can use TLS or DTLS to ensure communication security.

[0081] Information resources (video teaching clips, illustrated manuals) that provide first aid guidance are stored in CDN (content distribution network) nodes and can be called at any time via HTTPS.

[0082] Functional description:

[0083] When the ambulance has not yet arrived at the scene, the guidance module provides text, voice or video guidance to on-site personnel (such as correct cardiopulmonary resuscitation steps and hemostasis techniques), and obtains on-site feedback (changes in the patient's current status, whether the measures are implemented smoothly) through interactive questions and answers within the App, and transmits the feedback data back to the system.

[0084] Data analysis and model updating module

[0085] Hardware configuration:

[0086] Deployed on a dedicated data analysis server cluster (such as multiple Dell PowerEdge R740s running Hadoop and Spark clusters).

[0087] Software tools and protocols:

[0088] Use Spark or Flink for batch processing and streaming data analysis, and use HDFS or AWS S3 to store historical data.

[0089] The analysis results are written back to the parameter file of the optimization decision module or the model registry (such as MLflow) through the internal API (REST or gRPC).

[0090] Model update process:

[0091] After the event, the module summarizes and models the actual arrival time, the patient's final outcome, decision deviations, and execution difficulties, and extracts new parameters and rules through statistical learning and machine learning algorithms (such as XGBoost, Random Forest, and GraphNeural Network). The updated parameter files or trained new models are uploaded to the decision module and status evaluation module servers in a secure manner (HTTPS / SSH), thereby achieving performance improvement for the next rescue event.

[0092] Overall system communication and security

[0093] Network architecture:

[0094] The system uses Gigabit to 10 Gigabit Ethernet connections (10GbE / 25GbE) internally and connects to mobile terminals externally through 4G / 5G cellular networks or Wi-Fi6 networks.

[0095] Safety measures:

[0096] The entire data transmission process uses TLS encrypted channels, authorized access is controlled using OAuth 2.0 or JWT (JSON WebToken) token mechanism, and sensitive medical data complies with relevant privacy protection regulations (such as GB / T 35273 Personal Information Security Specification).

[0097] Terminal model example:

[0098] Command center server: Dell PowerEdge series

[0099] Ambulance GPS terminal: UniStrong UG902 or ZTE ZM710

[0100] Traffic data platform connection: Cisco ISR router (such as C1111 series)

[0101] Hospital HIS connection gateway: Huawei AR series enterprise gateway

[0102] On-site user smartphones: Huawei Mate 40, Apple iPhone 14 series, etc. (with dedicated APP installed)

[0103] Video call hardware: Huawei TE10 video conferencing terminal, Logitech C930e camera

[0104] Data analysis node: Inspur NF5280M5 deploys Spark cluster

[0105] Through the collaborative work of the above modules and terminal devices, the system of the present invention can achieve a complete closed loop from data acquisition, status assessment, optimized decision-making, dynamic execution adjustment, on-site guidance to post-analysis and model update in actual operation. The modules are interconnected through standardized interfaces and security protocols, and the transmission of real-time data, instructions and feedback is efficient and reliable. The emergency process is continuously iterated and optimized, ultimately achieving an effective reduction in the time from call to treatment and continuous improvement in the quality of rescue.

[0106] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the present invention. The specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for optimizing call-to-help time, characterized in that: The following steps are involved: (1) Obtain real-time data: including emergency vehicle location data, traffic status data, medical resource availability data, patient call information and condition description; (2) assess the urgency of the patient's condition and resource status; (3) Based on the above evaluation results and real-time data, an optimization algorithm is used to generate an ambulance dispatch and route planning plan; (4) Track the execution process in real time and dynamically adjust the scheduling plan according to environmental changes; (5) Provide first aid guidance to on-site personnel through remote communication before the arrival of rescue personnel; (6) Store and analyze the data and results of the entire process.

2. The method for optimizing the call-to-help time according to claim 1, characterized in that: The optimization algorithm is one or more of integer linear programming, genetic algorithm or deep reinforcement learning algorithm.

3. The method for optimizing the call-to-help time according to claim 1, characterized in that: The remote communication means include voice calls, video transmission, image displays and text prompts.

4. The method for optimizing the call-to-help time according to claim 1, characterized in that: The dynamic adjustment includes re-planning the route according to real-time traffic congestion information, reallocating ambulances according to resource shortages, and modifying the priority dispatch strategy according to patient status update information.

5. A call-to-help time optimization guidance system, comprising: Data collection and fusion module, used to obtain and integrate emergency vehicle location, traffic information, medical resource status, and patient call data; Status assessment module, used to assess the urgency of the patient's condition and the availability of rescue resources; Optimize the decision-making module to dispatch ambulances and plan routes based on the above evaluation results and data; Dynamic adjustment module, used to adjust the generated dispatch and route plans according to real-time change information; Remote guidance module, used to provide first aid guidance to the scene before the arrival of rescuers; The storage and recording module is used to store and archive relevant data, decision-making plans and processing results.

6. The call-to-treatment time optimization guidance system according to claim 5, characterized in that: The remote guidance module provides treatment suggestions to on-site personnel in the form of voice, video, images and text.

7. The call-to-treatment time optimization guidance system according to claim 5 or 6, characterized in that: The optimization decision module selects an applicable optimization algorithm to solve the scheduling plan based on preset optimization objectives, such as the shortest arrival time, the minimum resource occupancy or the maximum patient treatment rate.

Images