Intelligent infusion evaluation reminding method and system

Abstract

The invention relates to an intelligent infusion evaluation reminding method and system, and solves the problems that it is difficult to accurately grasp the illness state change and infusion requirements of each patient, and omission or delayed processing is likely to occur. The method comprises the steps that the environment noise value and the working state of the environment where medical staff currently works are obtained, performing fuzzification processing on the acquired environmental noise value and the working state of the medical staff, converting the environmental noise value and the working state of the medical staff into elements in a fuzzy set, and establishing a logical relationship among the environmental noise, the working state of the medical staff and the reminding level according to a pre-established fuzzy rule base; and generating corresponding fuzzy output results for different reminding levels through fuzzy reasoning, determining a specific reminding mode corresponding to the reminding levels through defuzzification calculation, and executing the corresponding specific reminding mode. The system has the advantages that efficient management and optimization of the infusion process are achieved, the working efficiency of medical staff is improved, and safety and effectiveness of infusion treatment are guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to an intelligent infusion assessment reminder method and system. Background Art

[0002] With the aging of the population and the continuous advancement of medical technology, the number of patients in hospitals is increasing, and the demand for medical care is also increasing. As a common treatment method, infusion occupies an important position in the daily medical work of hospitals.

[0003] Currently, during hospital infusion treatment, medical staff need to regularly inspect the patient's infusion status and manually record and adjust the parameters of the infusion pump.

[0004] In the above-mentioned infusion management, it mainly relies on manual observation and experience judgment of medical staff. This method has many limitations, such as it is difficult to accurately grasp the changes in the condition and infusion needs of each patient, and it is easy to miss or delay treatment. At the same time, manual management is inefficient and cannot meet the infusion management needs of a large number of patients. Summary of the invention

[0005] In order to achieve efficient management and optimization of the infusion process, improve the work efficiency of medical staff, and ensure the safety and effectiveness of infusion therapy, the present application provides an intelligent infusion assessment reminder method and system.

[0006] In the first aspect, the present application provides an intelligent infusion assessment reminder method, which adopts the following technical solution: An intelligent infusion assessment reminder method, comprising: Obtain information related to the patient's condition and infusion task entered into the system; Determine the score of the corresponding indicator according to the mapping relationship between the range of each indicator data in the disease-related information and the infusion task-related information and the score; According to the preset time decay formula, the weight of each indicator score changing over time is calculated according to the time elapsed since the indicator data was collected. At the same time, the gene polymorphism data and past medical history are extracted from the patient's condition-related information and input into the pre-built individual difference correction coefficient analysis model to output the correction coefficient for each indicator score; Combine the original scores of each indicator in the patient's condition-related information and the infusion task-related information, multiply them by the corresponding time decay weight and the individual difference correction coefficient, and then add them up to calculate the overall score of each patient. Automatically sort the patient's infusion data in descending order according to the overall score, and display them on the display interface of the mobile terminal held by the medical staff; According to the mapping relationship between the interval range of the overall score and the reminder level, the reminder level is analyzed and determined, wherein the reminder level is divided into three levels: high urgency, medium urgency, and low urgency; Obtain the ambient noise value and working status of the environment where the medical staff is currently working, perform fuzzy processing on the acquired ambient noise value and working status of the medical staff, convert them into elements in the fuzzy set, and establish the logical relationship between the ambient noise, working status of the medical staff and the reminder level based on the pre-established fuzzy rule base; Through fuzzy reasoning, corresponding fuzzy output results are generated for different reminder levels. Then, through defuzzification calculation, the specific reminder method corresponding to the reminder level is determined, and the corresponding specific reminder method is executed.

[0007] Optionally, after executing the corresponding specific reminder method, it also includes: Continuously analyze whether there are high-urgency alerts; If not, maintain the original setting; If yes, then obtain the scene information of the patient, and collect the working status, location information, workload of the medical staff, and the severity and urgency of the patient's condition in real time. The scene information includes the general ward scene, emergency rescue scene, operating room scene and emergency room scene; According to the preset quantitative scoring method, the working status, location information, workload, environmental noise value of medical staff, and the severity and urgency of the patient's condition are quantitatively scored; Based on the quantitative scoring of medical staff's work status score, location information score, workload score, and environmental noise value score, the preset optimal task allocation method based on particle swarm and ant colony algorithm is used to analyze, determine, and execute the optimal task allocation plan.

[0008] Optional optimal task allocation methods based on particle swarm and ant colony algorithms include: Randomly generate a preset number of N ants, each ant corresponds to a medical staff allocation plan, create a pheromone matrix with the number of rows equal to the number of medical staff M and the number of columns equal to 1, and the initial value is set to the preset constant T 0 , and synchronously generate a preset number P of particles, each particle represents a task allocation scheme, with preset initial position and speed parameters; The work status score, location information score, workload score, environmental noise value score of the medical staff obtained by quantitative scoring, as well as the severity score and urgency score of the patient's condition are input into the preset fitness calculation function. According to the pre-set weight adjustment rules for different scenarios of the patient, the corresponding weight is assigned to each scoring factor. The fitness calculation function multiplies each score with the corresponding weight and accumulates them to obtain the fitness value of the task allocation scheme represented by each ant and each particle; Particles update their positions and speeds according to the preset rules and the optimal positions of themselves and the particle group. Each iteration uses a preset function to calculate the fitness value. If the new value is better than the historical optimal value, the corresponding extreme value is updated. After the particle swarm finds the global optimal solution, it increases the pheromone concentration of the ant colony's high-quality paths according to the preset rules. When the ants choose a path, they incorporate the particle swarm's global optimal factors into the selection probability calculation according to specific weights. When the particles are updated, if the pheromone concentration in a certain area of ​​the ant colony exceeds the preset threshold, the probability of moving to that area is increased according to the preset rules. Ants select medical staff according to the preset path selection probability formula. The path selection probability formula integrates pheromone concentration, heuristic information and particle swarm global optimal factors. In different scenarios, the weights of each factor are dynamically adjusted according to the preset weight adjustment formula to achieve path selection. Each time the ants complete the path selection, the pheromone matrix will be updated according to the preset pheromone update formula. When the preset number of iterations T is reached or the fitness improvement in n consecutive iterations is less than the preset value, the ant colony iteration is stopped; According to the preset solution screening formula, the solutions represented by ants and particles are evaluated, and the solution with the highest fitness is selected; Assign high-emergency patient tasks to the medical staff corresponding to the plan, and send the patient's condition and scenario to the mobile terminal held by the medical staff through the hospital information channel according to the preset information push rules.

[0009] Optionally, after executing the corresponding specific reminder method, it also includes: Analyze whether the mobile terminal receives voice instructions from medical staff to view infusion-related information of the patient in charge; If yes, retrieve the patient's infusion related information according to the voice command and make a voice reply; If not, keep the original setting.

[0010] Optionally, the patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Within a preset time period, the voice commands received by the mobile terminal for viewing the infusion-related information of different patients are recorded, and the timestamp corresponding to each command is marked; According to the number of voice commands, the attention coefficient is preliminarily determined according to the corresponding relationship between the preset number range and the attention coefficient; According to the preset time stage weight adjustment rules, the initial attention coefficient is adjusted according to different time stages; According to the preset condition status weight adjustment method, the attention coefficient is adjusted again in combination with the patient's condition status included in the patient's condition related information; According to the preset multi-dimensional data influence rules, the fluctuation of vital signs within the preset time range contained in the patient's condition-related information and the frequency of changes in the infusion speed contained in the infusion task-related information are comprehensively considered to make a final correction to the attention coefficient; The determined attention coefficient was multiplied by each patient's overall score to obtain the final overall effective score; According to the final overall effective score, the patient infusion data is automatically sorted from high to low, and the sorting results are presented on the display interface of the mobile terminal held by the medical staff.

[0011] Optionally, a voice interaction step is also included, as follows: Analyze whether the infusion pump receives the patient's voice description information about the infusion feeling during the infusion, and define that the infusion pump is provided with the function of receiving the voice description information and transmitting it; If yes, the patient's voice description of the infusion feeling during the infusion is transmitted to the mobile terminal held by the medical staff, and the voice description information fed back by the medical staff is received and transmitted back to the infusion pump and the medical staff's voice description is made; If not, keep the original setting.

[0012] Optionally, after analyzing whether the infusion pump receives the patient's voice description information about the infusion feeling during the infusion, the following steps are also included: If it is confirmed that the infusion pump has received the patient's valid voice expression information, the patient's voice expression information is input into the pre-built bidirectional long short-term memory network keyword matching model, and the preset emergency keyword matching result is output, and the keyword urgency is graded to obtain the keyword matching score K; The patient's speech expression information is input into the medical speech sentiment analysis model based on convolutional neural network, and the speech sentiment intensity score S is output; The real-time location coordinates of the patient are obtained by using the UWB positioning system, and the area is determined by combining the hospital map. The hospital regional location evaluation model based on the hierarchical analysis method is used to input the emergency treatment resources of each area of ​​the hospital and the frequency of historical emergency events, output the regional priority weight, and calculate the location association score P; Extract the infusion start time from the infusion pump and related systems, analyze and obtain the infusion duration, input the infusion duration into the disease deterioration probability prediction model based on time series analysis, and output the adjusted time weight coefficient α; Using the emotion weight dynamic adjustment model based on machine learning, the historical data of the patient's voice emotion during the past infusion and the current severity of the disease are input, and the dynamically determined emotion intensity weight coefficient β is output; A location weight determination model based on the hierarchical analysis method is used to input the hospital area emergency treatment priority and patient location information, and output a reasonable location weight coefficient γ; Substitute the calculated α, K, β, S, γ, and P into the preset emergency assessment value calculation formula to obtain the emergency assessment value, where the preset emergency assessment value calculation formula is as follows: =α×K+β×S+γ×P; The patient's voice-expressed information is combined with an urgency assessment value , real-time location information and infusion-related parameters are packaged and transmitted to the medical staff’s mobile terminal.

[0013] Optionally, the patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Within a preset time period, the number of voice instructions or voice descriptions of medical staff and feedback to medical staff on infusion-related information for different patients is recorded through the voice interaction system of the mobile terminal. , and simultaneously use speech recognition and semantic analysis technology to identify the urgency-related keywords mentioned in each speech and mark the relevant speech; Count the frequency of emergency keywords in each patient's relevant speech ; The comprehensive attention factor of each patient is calculated using the preset comprehensive attention factor calculation formula. The comprehensive attention factor calculation formula is as follows: ; in, is the comprehensive attention factor, The number of voice commands The weight coefficient of The frequency of urgent keywords The weight coefficient of Assess the urgency level The weight coefficient of Multiply each patient's comprehensive concern factor by the patient's original overall score to obtain the overall effective score; According to the overall effective score, the patient infusion data is automatically sorted from high to low, and the sorting results are presented on the display interface of the mobile terminal held by the medical staff.

[0014] In the second aspect, the present application provides an intelligent infusion assessment reminder system, which adopts the following technical solution: An intelligent infusion assessment reminder system comprises a memory, a processor and a program stored in the memory and executable on the processor, wherein the program can implement the intelligent infusion assessment reminder method as described in the first aspect when loaded and executed by the processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of an intelligent infusion assessment reminder method according to an embodiment of the present application.

[0016] Figure 2 It is a flowchart diagram of the steps after executing the corresponding specific reminder method in another embodiment of the present application. DETAILED DESCRIPTION

[0017] The present application is further described in detail below in conjunction with the accompanying drawings.

[0018] Reference Figure 1 , is an intelligent infusion assessment reminder method disclosed in the present application, comprising: Step S100, obtaining patient condition related information and infusion task related information entered into the system.

[0019] Among them, the patient's condition-related information includes: the diagnosis details of the patient's current disease, symptoms, various physiological index values ​​(such as body temperature, blood pressure, heart rate, etc.), gene polymorphism data, and past medical history, etc. This information is used to comprehensively evaluate the patient's physical condition. The patient's condition-related information is obtained as follows: extract disease diagnosis, symptoms, past medical history and other information from the patient's electronic medical record through the hospital information management system (HIS); collect the patient's physiological index data in real time by connecting to various medical monitoring equipment in the ward (such as thermometers, sphygmomanometers, ECG monitors, etc.); gene polymorphism data comes from the previous special genetic testing laboratory report, which is entered into the system when the patient is admitted to the hospital.

[0020] Infusion task related information: including the type, dosage, infusion speed, estimated infusion duration and other information directly related to the infusion operation. The infusion task related information is obtained as follows: Before performing the infusion operation, the medical staff manually enters the type, dosage, infusion speed and other information of the infusion drug through the hospital's nursing information system, and the system automatically calculates the estimated infusion duration.

[0021] Step S200, determining the score of the corresponding indicator according to the mapping relationship between the range of each indicator data in the condition-related information and the infusion task-related information and the score.

[0022] Among them, indicator data refers to the specific data in the patient's condition-related information and infusion task-related information, such as body temperature and blood pressure in terms of the condition, and drug dosage and infusion speed in terms of the infusion task.

[0023] Score mapping relationship: the association rules between the pre-set indicator data range and the corresponding score, for example, a body temperature of 37.3-38℃ corresponds to a score of 2 points.

[0024] The general process is described as follows: the system compares the patient's condition and infusion task-related indicator data obtained in step S100 with the pre-set score mapping relationship, determines the score corresponding to the range of each indicator data, and provides a basis for the subsequent overall score calculation.

[0025] Step S300, according to the preset time decay formula, calculate the weight of each indicator score changing with time according to the time elapsed since the indicator data was collected, and at the same time extract genetic polymorphism data and past medical history from the patient's condition-related information, and input them into the pre-constructed individual difference correction coefficient analysis model, and output the correction coefficient for each indicator score.

[0026] Among them, the time decay formula: a mathematical expression used to reflect the change in the weight of the indicator score over time. The weight of the indicator score: the proportion of each indicator in the overall score will change over time. Gene polymorphism data: the variation information at a specific position in an individual's gene sequence will affect drug response and efficacy. Past medical history: a record of diseases a patient has had in the past. Individual difference correction coefficient analysis model: a model constructed through machine learning or statistical methods, which takes gene polymorphism data and past medical history as input and outputs a correction coefficient for the score of each indicator.

[0027] The general process is as follows: Calculate the time decay weight: Calculate the weight of each indicator score based on the time decay formula and the time from the indicator data collection time to the current time. For example, the time decay formula can be ,in, is the weight at time t, is the initial weight, is the attenuation coefficient, It is the time elapsed from the moment the indicator data is collected. Assuming that the initial weight of an indicator is =0.8, attenuation coefficient = 0.1, after 2 hours ( =2), then the weight calculation of this indicator is as follows: .

[0028] Obtaining individual difference correction coefficients: Extracting gene polymorphism data and past medical history from the patient's condition-related information, inputting the individual difference correction coefficient analysis model, and obtaining correction coefficients for each indicator score. For example, if the model outputs a correction coefficient of 1.2 for a certain indicator, it means that the score of the indicator needs to be multiplied by 1.2 for correction.

[0029] Step S400, combining the original scores of each indicator in the patient's condition-related information and the infusion task-related information, multiplying them by the corresponding time attenuation weight and the individual difference correction coefficient and then adding them up, so as to calculate the overall score of each patient, and automatically sorting the patient's infusion data in order from high to low according to the overall score, and displaying it on the display interface of the mobile terminal held by the medical staff.

[0030] Among them, the original score: the initial score of each indicator in the patient's condition-related information and infusion task-related information determined in step S200 based on the indicator data range and the score mapping relationship. Time decay weight: the weight of each indicator score that changes with time is calculated in step S300 according to the time decay formula based on the elapsed time since the indicator data was collected. Individual difference correction coefficient: the correction coefficient for each indicator score output after step S300 inputs the patient's genetic polymorphism data and past medical history into the pre-built individual difference correction coefficient analysis model. Overall score: a score that reflects the comprehensive situation of the patient's infusion, obtained through a specific calculation method by comprehensively considering the original score of each indicator, the time decay weight and the individual difference correction coefficient.

[0031] Calculate the overall score: The system multiplies the original score of each indicator in the patient's condition-related information and infusion task-related information by the corresponding time decay weight and individual difference correction coefficient, and then adds them up. Assume that the patient has three indicators A, B, and C, and the original scores are SA=5, SB=4, and SC=3, respectively. The corresponding time decay weights are WA=0.8, WB=0.9, and WC=0.7, and the individual difference correction coefficients are CA=1.1, CB=1.2, and CC=1.3, respectively. The overall score calculation formula is S=SA×WA×CA+SB×WB×CB+SC×WC×CC. Then the patient's overall score S=11.45.

[0032] Step S500: analyzing and determining the reminder level according to the mapping relationship between the interval range that the overall score falls into and the reminder level, wherein the reminder level is divided into three levels: high urgency, medium urgency, and low urgency.

[0033] Among them, the reminder level: is used to intuitively indicate the urgency of the patient's infusion, which is divided into three levels: high urgency, medium urgency, and low urgency. Mapping relationship: the correspondence between the pre-set overall score range and the reminder level.

[0034] The general process is described as follows: After the system obtains the overall score of each patient from step S400, it compares it with the pre-set mapping relationship. For example, an overall score greater than 80 points corresponds to a high emergency reminder level, 50 to 80 points corresponds to a medium emergency reminder level, and less than 50 points corresponds to a low emergency reminder level. If a patient's overall score is 70 points, by comparing the mapping relationship, the system determines that the patient's reminder level is medium emergency. In this way, the system analyzes and determines the reminder level corresponding to each patient's infusion situation according to the interval range in which the overall score falls, providing a basis for subsequent reminders to medical staff.

[0035] Step S600, obtain the environmental noise value and working status of the environment in which the medical staff currently works, fuzzify the obtained environmental noise value and working status of the medical staff, convert them into elements in the fuzzy set, and establish a logical relationship between the environmental noise, the working status of the medical staff and the reminder level based on a pre-established fuzzy rule library.

[0036] Among them, the environmental noise value: obtained through the noise sensor arranged in the working environment of the medical staff, the sensor will transmit the collected environmental noise value to the system in real time. The working status of medical staff: manually input by medical staff through the special status selection interface on the mobile terminal, such as selecting "mildly busy" or "extremely busy" when busy; it can also be automatically determined by the system by analyzing the operation frequency of the mobile terminal, work task allocation and other data. Fuzzy rule base: jointly discussed and formulated by medical information experts and medical staff, and stored in the system.

[0037] The general process is described as follows: Data collection and fuzzification: The system obtains the environmental noise value and the working status of medical staff. For the environmental noise value, for example, the collected value is 65dB, and it is converted into an element in the fuzzy set, such as "moderately noisy" according to the pre-set fuzzification rules; for the working status of medical staff, if the medical staff manually selects "mildly busy", the status is converted into the corresponding fuzzy element.

[0038] Establishing logical relationships: The system establishes logical relationships between environmental noise, medical staff working status, and reminder levels based on the fuzzy rule base. For example, the fuzzy rule base may contain such a rule: when the environmental noise is "moderately noisy" and the medical staff working status is "mildly busy", the weight of the corresponding medium emergency reminder level increases. Although no specific formula is involved here, fuzzy rules can be understood as a logical relationship based on conditional judgment. A fuzzy rule base is formed by a large number of similar rules to achieve the association between the three.

[0039] Step S700, through fuzzy reasoning, corresponding fuzzy output results are generated for different reminder levels, and then through defuzzification calculation, the specific reminder method corresponding to the reminder level is determined, and the corresponding specific reminder method is executed.

[0040] Among them, fuzzy reasoning: based on fuzzy logic, according to the fuzzy rule base and fuzzified input information (such as environmental noise, working status of medical staff), the process of reasoning about the reminder level to obtain the corresponding fuzzy output result. Fuzzy output result: after fuzzy reasoning, the fuzzy conclusions generated for different reminder levels, for example, for a high emergency reminder level, the fuzzy output result may be a fuzzy expression of "very strong reminder". Defuzzification calculation: the process of converting the fuzzy output result into a specific and accurate reminder method, so that the system can perform the actual reminder operation.

[0041] The general process is described as follows: Fuzzy reasoning: The system uses the fuzzified environmental noise and medical staff work status information obtained in step S600 as input, and performs fuzzy reasoning in combination with the fuzzy rule base. For example, there is a rule in the fuzzy rule base that "if the environmental noise is 'noisy' and the medical staff work status is 'busy', then the fuzzy output of the high emergency reminder level is 'emergency alarm'". When the input meets this condition, the system obtains the fuzzy output result of "emergency alarm" for the high emergency reminder level. Defuzzification calculation: For fuzzy output results, the system converts them into specific reminder methods through defuzzification calculation. Common defuzzification methods include the maximum membership method. Assume that the fuzzy output result for the high emergency reminder level is a fuzzy set about the reminder intensity, where the membership function corresponding to "emergency alarm" is , find the x value with the largest membership by calculation. If the specific reminder method corresponding to this value is "continuously emit a high-decibel alarm sound", the system will execute this reminder method. In actual applications, the defuzzification calculation will accurately convert the fuzzy results into executable reminder actions based on a variety of pre-set defuzzification algorithms and mapping relationships, so as to execute the corresponding specific reminder methods for different reminder levels and promptly inform medical staff of the emergency situation of patient infusion.

[0042] Reference Figure 2 , an intelligent infusion assessment reminder method further includes steps after executing the corresponding specific reminder method, which are as follows: Step S800, continuously analyzing whether a high-urgency reminder occurs. If not, executing step S900; if yes, executing step SA00.

[0043] High urgency reminder: determined in step S500, indicating that the patient's infusion situation is extremely urgent and requires immediate treatment by medical staff, and the corresponding overall score falls into the high urgency interval range. Continuous analysis: The system continuously monitors and judges the current alert levels to determine whether there are any high-urgency alerts.

[0044] Step S900, maintaining the original settings.

[0045] Original settings: covers a series of parameters, configurations, and task allocation strategies maintained by the intelligent infusion assessment reminder system before executing a specific reminder method. These settings include but are not limited to the default adjustment rules for infusion speed, reminder frequency settings, the established mode of medical staff task allocation, and the evaluation model parameters determined based on the patient's condition and infusion task-related information.

[0046] Step SA00, obtain the scene information of the patient, and collect and obtain the working status, location information, workload, and severity and urgency of the patient's condition of the medical staff in real time.

[0047] Among them, the patient's scene information: used to indicate the patient's current medical environment, including general ward scenes, emergency rescue scenes, operating room scenes, and emergency room scenes. Different scenes have different requirements for the allocation and response of medical resources. The patient's scene information can be obtained through the records of the patient's ward, treatment area and other information in the hospital information system; medical staff can also manually enter the scene-related information when the patient registers for admission.

[0048] Medical staff work status: describes the current busyness and concentration of medical staff, such as idle, slightly busy, extremely busy, etc., which can reflect the energy that medical staff can devote to new tasks. The work status of medical staff can be obtained through a dedicated mobile application, manually selecting their own work status to update; or the system can automatically analyze the frequency of medical staff operating mobile terminals, task processing time and other data.

[0049] Medical staff location information: refers to the specific geographical location of medical staff in the hospital. It is obtained through positioning technology, which helps to quickly deploy personnel to where they are needed. The location information of medical staff can be obtained by using indoor positioning systems deployed by hospitals, such as Bluetooth positioning, Wi-Fi positioning and other technologies, to obtain real-time signals from positioning devices (such as work mobile phones, badges, etc.) carried by medical staff to determine their location.

[0050] Medical staff workload: reflects the number of tasks and pressure currently undertaken by medical staff, which can be comprehensively evaluated by counting the number of patients they are responsible for, the urgency of tasks, etc. The workload of medical staff can be obtained by counting the number of patients they are responsible for, the estimated time required for various tasks, etc. in the hospital nursing management system, and the workload can be calculated.

[0051] Severity of the patient's condition: An assessment of the severity of the patient's condition based on multiple factors, including the type of disease, symptoms, and degree of abnormality of physiological indicators. The severity of the patient's condition can be determined by analyzing information such as disease diagnosis, examination and test reports, and real-time physiological monitoring data in the patient's electronic medical record, based on a pre-set condition assessment model.

[0052] Patient urgency: Determine the urgency of the patient's need for immediate medical intervention based on the patient's current infusion status, changes in condition, etc. The patient's urgency can be determined based on the reminder level determined in step S500, the patient's real-time infusion data, and changes in condition.

[0053] The general process is described as follows: When the system determines in step S800 that a high-urgency reminder appears, step SA00 is executed. The system first obtains the patient's scene information from the above-mentioned corresponding information source. For example, through the query of the hospital information system, it is known that a high-urgency patient is in a general ward scene. At the same time, the corresponding technical means are used to obtain the working status, location information and workload of medical staff. Suppose through feedback from a mobile application, it is known that a medical staff is in a slightly busy state. With the help of the positioning system, its location is determined to be at the nurse station on the third floor of the hospital. Its workload is calculated from the nursing management system to be at a medium level. The severity and urgency of the patient's condition are obtained according to the established evaluation method and data source. This step is mainly an information collection process, which does not involve complex formula calculations. It aims to provide a comprehensive data basis for subsequent quantitative scoring and task allocation.

[0054] Step SB00, according to the preset quantitative scoring method, quantitative scoring is performed on the medical staff's working status, location information, workload, environmental noise value, and the severity and urgency of the patient's condition.

[0055] Among them, the preset quantitative scoring method is a set of pre-established rules used to convert factors such as the working status of medical staff, location information, workload, environmental noise value, and the severity and urgency of the patient's condition into specific numerical scores for subsequent task allocation analysis.

[0056] Quantitative scoring: A specific score obtained by converting various non-numerical information (such as working status, scene, etc.) or numerical information (such as environmental noise value) according to preset rules.

[0057] Assume that the quantitative scoring method is as follows: The working status score of medical staff was as follows: idle was 10 points, slightly busy was 6 points, moderately busy was 3 points, and extremely busy was 1 point.

[0058] Medical staff location information score: 10 points if the distance from the patient is within 10 meters, 6 points if the distance from the patient is between 10 and 30 meters, 3 points if the distance from the patient is between 30 and 50 meters, and 1 point if the distance from the patient is more than 50 meters.

[0059] Workload score of medical staff: low workload is 10 points, medium workload is 6 points, high workload is 3 points, and extremely high workload is 1 point.

[0060] Environmental noise value score: Noise value below 40dB is 10 points, 40-60dB is 6 points, 60-80dB is 3 points, and over 80dB is 1 point.

[0061] The severity of the patient's condition was scored as follows: 1 point for mild, 3 points for general, 6 points for severe, and 10 points for critical.

[0062] Patient urgency score: low urgency is 1 point, medium urgency is 3 points, and high urgency is 10 points.

[0063] If a medical staff member is slightly busy at work, is 20 meters away from the patient, has a medium workload, and has an ambient noise value of 50dB, the corresponding patient's condition is severe and the urgency is high. The medical staff member's work status score is 6 points, the location information score is 6 points, the workload score is 6 points, the ambient noise value score is 6 points, the patient's condition severity score is 6 points, and the patient's urgency score is 10 points.

[0064] Step SC00, based on the medical staff's work status score, location information score, workload score, and environmental noise value score obtained by quantitative scoring, a preset optimal task allocation method based on particle swarm and ant colony algorithm is used to analyze and determine the optimal task allocation plan and execute it.

[0065] The scores obtained by quantitative scoring are: that is, the specific scores obtained for the working status, location information, workload, environmental noise value of medical staff, and the severity and urgency of the patient's condition in step SB00 according to the preset quantitative scoring method. These scores quantify the impact of relevant factors on task allocation.

[0066] Preset optimal task allocation method based on particle swarm and ant colony algorithm: an intelligent optimization algorithm that combines particle swarm algorithm and ant colony algorithm. The particle swarm algorithm simulates the foraging behavior of bird flocks, allowing particles (representing possible solutions) to iterate in the solution space to find the optimal solution; the ant colony algorithm simulates the mechanism of ants releasing pheromones during foraging, and uses the pheromone concentration to guide ants to find the optimal path. In task allocation, the path corresponds to the task allocation plan. This composite algorithm is used to find the optimal task allocation plan. For details, please refer to steps SC01 to SC07.

[0067] The optimal task allocation methods based on particle swarm and ant colony algorithms include: Step SC01, randomly generate a preset number N of ants, each ant corresponds to a medical staff allocation plan, create a pheromone matrix with the number of rows equal to the number of medical staff M and the number of columns equal to 1, and the initial value is set to a preset constant T 0 , and synchronously generate a preset number P of particles, each particle represents a task allocation scheme, with preset initial position and speed parameters.

[0068] Ants: In the optimal task allocation method based on particle swarm and ant colony algorithm, ants are the carriers of the simulated ant colony algorithm. Each ant corresponds to a medical staff allocation plan. By simulating the selection behavior of ants on different "paths" (that is, different medical staff allocation plans), the optimal task allocation plan is found.

[0069] Pheromone matrix: A matrix with M rows and 1 columns, which is used to store the pheromone concentration corresponding to each possibility of medical staff allocation. The pheromone concentration will be updated with the iteration of the algorithm, affecting the probability of ants choosing a path (i.e., medical staff allocation plan).

[0070] Preset constant T 0 : The initial value of the pheromone matrix is ​​a pre-set fixed value, which serves as the initial state of the pheromone concentration when the algorithm starts. Particles: In the particle swarm algorithm, particles are individuals that represent possible solutions, and each particle represents a task allocation scheme. Particles search for the optimal solution in the solution space by constantly adjusting their position and speed. Preset numbers N and P: N is the preset number of randomly generated ants, and P is the preset number of synchronously generated particles. These numbers are parameters pre-set by the algorithm developer based on the scale of the actual problem and computing resources to control the search range and computational complexity of the algorithm. Initial position and velocity parameters: The initial position and velocity of the particle at the beginning of the search for the optimal solution are pre-set parameters that determine the starting state of the particle search. For example, the initial position can be represented by a vector, and each element in the vector corresponds to a dimension in the solution space; the initial velocity is also represented by a vector, which determines the direction and speed of the particle's movement at the beginning.

[0071] The general process is described as follows: When the optimal task allocation method based on particle swarm and ant colony algorithm is started, the system first randomly generates ants according to the preset number N. Each ant is assigned an initial medical staff allocation plan. For example, assuming there are 3 medical staff (numbered 1, 2, and 3), the initial allocation plan for an ant may be that medical staff 1 is responsible for patient A, medical staff 2 is responsible for patient B, and medical staff 3 is responsible for patient C. At the same time, a pheromone matrix with the number of rows equal to the number of medical staff M (assuming M=3) and the number of columns equal to 1 is created. The initial value of each element of the matrix is ​​set to the preset constant T0 (assuming T0=0.1), that is, the pheromone matrix is In addition, the system synchronously generates particles according to the preset number P. Each particle represents a task allocation scheme, and sets the initial position and speed parameters for the particle. Assuming that two particles are generated (P=2), the solution space is three-dimensional, and the initial position of particle 1 can be set to (represents a task allocation scheme), the initial speed is set to ; The initial position and speed of particle 2 are also set in a similar way. Through this series of operations, the foundation is laid for the subsequent use of particle swarm and ant colony algorithms to search for the optimal task allocation solution.

[0072] Step SC02, the medical staff work status score, location information score, workload score, environmental noise value score, and patient condition severity score and urgency score obtained by quantitative scoring are input into the preset fitness calculation function, and each scoring factor is assigned a corresponding weight according to the pre-set weight adjustment rules for different scenarios in which the patient is located. The fitness calculation function multiplies each score with the corresponding weight and accumulates them to obtain the fitness value of the task allocation scheme represented by each ant and each particle.

[0073] Among them, the fitness calculation function is a function used to evaluate the quality of the task allocation scheme represented by each ant and each particle. By multiplying each scoring factor with the corresponding weight and accumulating them, a value is obtained. The larger the value, the better the scheme. For example, assuming that the fitness calculation function can be as follows: ;in, They are the medical staff working status score, location information score, workload score, environmental noise value score, patient condition severity score and urgency score. is the corresponding weight.

[0074] Weight adjustment rules: According to the different scenarios in which the patients are located, different weights are assigned to each scoring factor (medical staff work status score, location information score, etc.) to reflect the differences in the importance of each factor in different scenarios.

[0075] Fitness value: A value calculated by the fitness calculation function, used to measure the quality of each task allocation plan.

[0076] The general process is as follows: the system obtains the medical staff work status score, location information score, workload score, environmental noise value score, patient condition severity score and urgency score obtained in step SB00 from the data storage module. At the same time, according to the patient's scenario, the corresponding weight adjustment rules are read from the configuration file to assign corresponding weights to each scoring factor.

[0077] In step SC03, the particle updates its position and speed according to the preset rules and the optimal position of itself and the particle group. The fitness value is calculated using the preset function in each iteration. If the new value is better than the historical optimal value, the corresponding extreme value is updated.

[0078] Among them, preset rules: In the particle swarm algorithm, the established rules for updating the position and velocity of particles are usually based on the particle's own historical optimal position and the global optimal position of the particle swarm.

[0079] The optimal position of the particle itself: the position with the best fitness value experienced by the particle during the iteration process, which represents the local optimal task allocation solution searched by the particle.

[0080] Global optimal position of particle swarm: The position with the best fitness value found by all particles in the particle swarm during the iteration process represents the global optimal task allocation solution searched by the entire particle swarm.

[0081] Preset function: a function used to calculate the fitness value of the particle at the new position, which is the same as the fitness calculation function in step SC02.

[0082] Historical optimal extreme value: the optimal fitness value recorded by the particle itself or the particle group in the previous iteration process.

[0083] The general process is described as follows: During the iteration process of the particle swarm algorithm, each particle updates its position and speed according to the preset rules. The common speed and position update formulas are as follows: The speed update formula is as follows: .

[0084] The position update formula is as follows: .

[0085] in, is the velocity of particle i at iteration t, is the position of particle i at iteration t, w is the inertia weight, and is the learning factor, and is a random number between 0 and 1. is the optimal position of particle i itself, is the global optimal position of the particle swarm.

[0086] Each time a particle updates its position, the system uses a preset function to calculate the fitness value of the particle at the new position. If the new fitness value is better than the particle's historical optimal extreme value, the particle's own optimal position and historical optimal extreme value are updated.

[0087] Step SC04: After the particle swarm finds the global optimal solution, it increases the pheromone concentration of the ant colony's high-quality path according to the preset rules. When the ants select the path, they incorporate the particle swarm's global optimal factor into the selection probability calculation according to a specific weight. When the particles are updated, if the pheromone concentration in a certain area of ​​the ant colony exceeds the preset threshold, the probability of moving to the area is increased according to the preset rules.

[0088] Among them, the particle swarm global optimal solution: the task allocation solution with the highest fitness value found by the particle swarm during the iteration process of the particle swarm algorithm. Ant colony quality path: In the ant colony algorithm, the path corresponding to the medical staff allocation solution with a higher fitness value has a relatively high pheromone concentration. Pheromone concentration: indicates the quality of a certain path (i.e., the task allocation solution). The higher the pheromone concentration, the greater the probability that the ant will choose this path. Preset rules: established rules for adjusting pheromone concentration and particle movement probability, designed by the algorithm developer based on the algorithm principles and problem characteristics. Preset threshold: a critical value used to determine whether the pheromone concentration in a certain area of ​​the ant colony is too high. When this threshold is exceeded, it will affect the particle movement probability.

[0089] The general process is as follows: When the particle swarm finds the global optimal solution, the system increases the pheromone concentration on the high-quality paths of the ant colony according to the preset rules. This step is to guide the ants to choose more paths similar to the particle swarm's global optimal solution.

[0090] Assume that the pheromone update rule is: if the similarity between a certain path and the global optimal solution of the particle swarm reaches a certain level, the pheromone concentration of the path increases .

[0091] When ants choose paths, the calculation of path selection probability will be integrated into the particle swarm global optimal factor. For example, the path selection probability formula may be as follows: ; in, is the probability that an ant chooses a path from position i to position j, is the path at time t The pheromone concentration, is heuristic information (such as the inverse of the distance), and is a parameter that controls the relative importance of pheromone and heuristic information, is the set of next positions that the ant can choose when it is at position i, It is a factor related to the global optimal solution of the particle swarm.

[0092] When the particle is updated, if the pheromone concentration in a certain area of ​​the ant colony exceeds the preset threshold, the particle will increase the probability of moving to that area according to the preset rules. Exceeding the preset threshold , the probability of the particle moving to this area increases .

[0093] In this way, the particle swarm algorithm and the ant colony algorithm cooperate with each other to find a better task allocation solution.

[0094] In step SC05, ants select medical staff according to the preset path selection probability formula. The path selection probability formula integrates pheromone concentration, heuristic information and the global optimal factor of the particle swarm. In different scenarios, the weights of each factor are dynamically adjusted according to the preset weight adjustment formula to achieve path selection. Each time the ants complete the path selection, they will update the pheromone matrix according to the preset pheromone update formula. When the preset number of iterations T is reached or the fitness improvement in n consecutive iterations is less than the preset value, the ant colony iteration is stopped.

[0095] Among them, the path selection probability formula is a mathematical formula used to calculate the probability of ants choosing different medical staff allocation paths (i.e., task allocation plans). It comprehensively considers pheromone concentration, heuristic information, and the global optimal factors of the particle swarm to determine the probability of ants choosing a path among many possible paths. Pheromone concentration: Pheromone is a concept used in simulating ant colony behavior. The pheromone concentration represents the frequency with which past ants have chosen a certain path and the quality of that path. In the task allocation scenario, the higher the pheromone concentration corresponding to a medical staff allocation plan, the more "favored" the plan is by past ants, and the greater the probability that it will be chosen by current ants. Heuristic information: A type of auxiliary information designed based on the characteristics of the problem itself, which can guide ants to make better path choices. For example, in the allocation of medical staff tasks, the distance between the medical staff and the patient, the degree of match between the medical staff's professional skills and the patient's condition, etc. can be used as heuristic information. The path corresponding to the medical staff who is close to the patient and has a high degree of professional skill match has a relatively high heuristic information value. Particle swarm global optimal factors: The relevant features or influencing factors corresponding to the global optimal solution found by the particle swarm algorithm during the search process. Incorporating them into the path selection probability calculation is to allow the ant colony algorithm to learn from the high-quality solutions found by the particle swarm algorithm, thereby accelerating the convergence to the global optimal task allocation solution. Weight adjustment formula: A formula that dynamically adjusts the weights of each factor (pheromone concentration, heuristic information, and particle swarm global optimal factor) in the path selection probability formula according to the different scenarios the patient is in. In different scenarios, the importance of each factor to task allocation is different, and the weight adjustment formula can flexibly change the influence of each factor. Pheromone update formula: After the ant completes a path selection (i.e. determines a medical staff allocation plan), the formula used to update the pheromone concentration of the corresponding path in the pheromone matrix. By updating the pheromone concentration, the path selection situation is recorded, providing a reference for subsequent ant path selection.

[0096] Assume that the pheromone update formula can be as follows: ; in, Is the updated path The pheromone concentration, is the pheromone volatility coefficient, Is the path after this path selection The increment of pheromone concentration (related to factors such as the fitness value of the path). The system calculates the increment of pheromone concentration based on the path selected by the ant. , update the pheromone concentration of the corresponding path in the pheromone matrix, and provide a basis for the next ant path selection.

[0097] Step SC06, according to the preset solution screening formula, the solutions represented by the ants and particles are evaluated, and the solution with the highest fitness is selected.

[0098] Among them, the scheme screening formula is a pre-set mathematical expression used to evaluate the quality of the task allocation schemes represented by ants and particles. The evaluation value of each scheme is calculated through this formula, and the scheme with the highest fitness is screened out based on this.

[0099] The general process is described as follows: the system retrieves all the task allocation schemes represented by ants and particles and their fitness values ​​from the data storage module. Then, each scheme is evaluated and calculated according to the scheme screening formula. Assume that the scheme screening formula is E=a×F+b×S, where E is the scheme evaluation value, F is the fitness value calculated in step SC02, S is the stability index of the scheme (for example, the balance of task allocation of medical staff in the scheme, etc., which can be calculated based on data such as the workload of medical staff), and a and b are weight coefficients, which are set by the algorithm developer according to actual conditions. For each scheme, the system substitutes its fitness value F and the calculated stability index S into the formula to obtain the corresponding evaluation value E. By comparing the evaluation values ​​E of all schemes, find the scheme with the largest evaluation value, which is the scheme with the highest fitness.

[0100] Step SC07, assign the high-emergency patient task to the medical staff corresponding to the plan, and send the patient's condition and scenario to the mobile terminal held by the medical staff through the hospital information channel according to the preset information push rules.

[0101] Among them, high-emergency patient tasks refer to infusion-related processing tasks for patients at a high-emergency alert level, including but not limited to emergency adjustment of infusion speed, inspection of infusion equipment, and observation of changes in the patient's condition. Hospital information channels: various systems and networks used for information transmission within the hospital, such as the hospital information management system (HIS), nursing information system, instant messaging platform, etc., through which patient-related information can be pushed to medical staff. Information push rules: pre-established rules that determine how, when, and to which medical staff members patient information is pushed, such as pushing information based on the ward area the medical staff is responsible for, professional skills matching, and other criteria.

[0102] The general process is described as follows: the system first obtains detailed information on high-emergency patient tasks from the task management module. Then, according to the optimal task allocation plan determined in step SC06, the medical staff responsible for the high-emergency patient task is identified. Then, through the hospital information channel, according to the preset information push rules, the patient's condition (including disease diagnosis, symptom manifestations, real-time physiological indicators, etc.) and the scene (ordinary ward scene, emergency rescue scene, etc.) are sent to the mobile terminal held by the corresponding medical staff. For example, if the information push rule is set to push according to the ward area responsible for the medical staff, and a high-emergency patient is in Ward 3, and the ward area is under the responsibility of medical staff A, the system will send the patient information to the mobile terminal device (such as a work phone, tablet computer) of medical staff A through the hospital information management system or the nursing information system.

[0103] Furthermore, an intelligent infusion assessment reminder method further includes steps after analyzing and determining the optimal task allocation plan and executing it, which are as follows: The patient's emergency assessment value is calculated at intervals of 30 minutes according to the preset emergency assessment value calculation formula. If the patient's emergency assessment value decreases in three consecutive assessments and the final E is <80 points (the high emergency score range is [80,100], and the medium emergency score range is [50,79]), the high emergency medical resource optimization allocation process is initiated.

[0104] The infusion drug information was obtained from the hospital information system, the infusion pump collected the infusion speed data, the medical monitoring equipment collected the vital sign change data, and the electronic medical record system extracted the historical condition data. Statistical methods were used to remove outliers and the Z-Score formula was used to standardize the data.

[0105] The drug risk level is divided according to the side effects and allergic risks of the drugs (3 points for high risk, 2 points for medium risk, and 1 point for low risk); the values ​​are quantified according to the infusion speed range (1 point for slow speed <10 drops / minute, 2 points for medium speed 10-30 drops / minute, and 3 points for fast speed >30 drops / minute), the degree of deviation of vital signs from the normal range (1 point for mild deviation, 2 points for moderate deviation, and 3 points for severe deviation), and the severity of the disease (1 point for mild, 2 points for moderate, and 3 points for severe).

[0106] The multivariate linear regression algorithm was used to build a model with infusion speed, drug dosage, patient age, and underlying diseases as independent variables and changes in vital signs as dependent variables. The historical infusion data was used for training and learning to predict the risk probability of adverse events. The DBSCAN clustering algorithm was used to take drug risk level, infusion speed score, vital sign deviation score, and adverse event risk probability as inputs to divide the infusion situation into stable infusion, potential risk infusion, and high-risk infusion categories.

[0107] Factors such as the infusion risk level, medical staff ability score (emergency treatment, adverse event handling experience, professional skills matching, etc.), and medical staff and patient position matching are included in the genetic algorithm fitness function. For high-risk infusion patients, the weight of the medical staff's emergency treatment ability score is set to ≥40%, and the weight of the infusion-related adverse event handling experience score is set to ≥30%. Personnel are randomly selected from the medical staff database to form the initial population, ensuring that the proportion of medical staff with experience in handling high-risk infusions is ≥20, and each individual represents a medical staff allocation plan.

[0108] Through selection (based on fitness value, such as roulette selection method), crossover (exchanging some medical staff allocation information), and mutation (high-risk infusion patients give priority to exchange or randomly allocate medical staff with relevant experience), the medical staff allocation plan is continuously iterated and optimized to improve the fitness value of the population. When the maximum number of iterations of the genetic algorithm set in advance is reached, the iteration is stopped and it is considered that the optimal medical staff allocation plan is obtained.

[0109] Step SC00, based on the medical staff's work status score, location information score, workload score, and environmental noise value score obtained by quantitative scoring, a preset optimal task allocation method based on particle swarm and ant colony algorithm is used to analyze and determine the optimal task allocation plan and execute it.

[0110] An intelligent infusion assessment reminder method further includes steps after executing the corresponding specific reminder method, which are as follows: Step Sa00, analyzing whether the mobile terminal receives a voice instruction from a medical staff member to view the infusion-related information of the patient in charge. If yes, execute step Sb00; if no, execute step Sc00.

[0111] Among them, mobile terminals refer to devices that medical staff carry with them for work, such as smart phones, tablets, etc. These devices are installed with the application of the intelligent infusion assessment reminder system and have the function of receiving voice commands and displaying information. Voice command: A command issued by medical staff to a mobile terminal by speaking to request to view infusion-related information of the patient they are responsible for, such as "view infusion information of [patient name]".

[0112] Step Sb00, retrieve the patient's infusion related information according to the voice command and make a voice reply.

[0113] Among them, the patient's infusion-related information covers: the patient's infusion progress (amount infused, remaining infusion volume), infusion speed, the name and dosage of the infused drug, the expected completion time of the infusion, and may also include the patient's current vital signs (such as body temperature, blood pressure and other physiological indicators related to infusion).

[0114] Voice reply: The system will provide voice feedback on the patient's infusion-related information to the medical staff who issued the voice command, making it convenient for medical staff to obtain information without checking the screen when they are busy.

[0115] Step Sc00, maintain the original settings.

[0116] The original settings refer to the current operating status settings of the intelligent infusion assessment reminder system when it does not receive voice commands from medical staff to view the infusion-related information of the patient it is responsible for. Including but not limited to reminder rule settings (such as reminder time intervals, reminder method selection), data display format settings (such as infusion progress displayed as a percentage or specific milliliters), and interactive configurations of the system and other medical devices or information systems. These settings ensure that the system operates according to established rules in daily workflows and provides medical staff with stable and continuous infusion assessment and reminder services.

[0117] The patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Step S401, within a preset time period, recording the voice commands received by the mobile terminal for viewing the infusion-related information of different patients, and marking the timestamp corresponding to each command.

[0118] Among them, the preset time period: a pre-set time period for counting voice commands, such as 1 hour, 1 day, etc. The time period is set to analyze the attention of medical staff to the infusion information of different patients within a relatively reasonable time range. Mobile terminal: A portable device used by medical staff to receive and process work-related information, such as smart phones, tablet computers, etc., with a built-in application for the intelligent infusion assessment reminder system and the ability to receive and recognize voice commands. Voice commands: Medical staff use the microphone of the mobile terminal to input voice expressions to query the infusion-related information of different patients, such as "check the infusion information of [patient name]". Timestamp: A mark that records the specific time when the voice command is received, accurate to seconds or even milliseconds, and is used for subsequent analysis of the time sequence and time interval of command reception.

[0119] The general process is as follows: During the operation of the system, the intelligent infusion assessment reminder system application on the mobile terminal works according to the preset time period (assuming the preset time period is 1 day). From the beginning of the day, the voice recognition module of the application is in a continuous monitoring state. When the medical staff issues a voice command to view the patient's infusion-related information, the voice recognition module converts the voice into text information, and at the same time records the exact time when the command is received, generating a corresponding timestamp.

[0120] Step S402, preliminarily determining the attention coefficient according to the number of voice instructions and the corresponding relationship between the preset number range and the attention coefficient.

[0121] The number of voice commands is the total number of voice commands received by the mobile terminal from the medical staff to check the infusion-related information of different patients within the preset time period of step S401. For example, in one day, the medical staff issued 20 voice commands to check the infusion information of the patient, and these 20 times are the number of voice commands. The correspondence between the preset frequency range and the attention coefficient: a set of pre-established rules is used to preliminarily determine the attention coefficient based on the number of voice commands. The attention coefficient reflects the degree of attention of medical staff to the patient's infusion information. Different voice command frequency ranges correspond to different attention coefficient values. For example, if the number of voice commands is 0-5 times, the attention coefficient is 0.5; 6-10 times, the attention coefficient is 0.7, etc. Attention coefficient: A value used to quantify the degree of attention paid by medical staff to patient infusion information. This value will be further adjusted in combination with other factors in subsequent steps and ultimately used to calculate the patient's overall effective score to more accurately reflect the priority of the patient's infusion situation.

[0122] Step S403, adjusting the preliminary attention coefficient according to different time stages according to the preset time stage weight adjustment rule.

[0123] Among them, the preset time stage weight adjustment rule is a pre-set rule for adjusting the initial attention coefficient according to different time stages. In different time stages, such as busy hours during the day and rest periods at night, the medical staff may have different attention needs for patient infusion information. This rule is based on this to adjust the attention coefficient to better fit the actual situation.

[0124] Time phase: Divide a day or other preset time period into different time periods, each of which has different characteristics, such as work intensity, the changing pattern of the patient's condition, etc. Common divisions include daytime (8:00-18:00), nighttime (18:00-8:00 the next day), etc.

[0125] Preliminary attention coefficient: In step S402, a value is preliminarily determined based on the correspondence between the number of voice instructions and the preset number range and the attention coefficient. This value will be further adjusted in step S403 based on the time stage factor.

[0126] The general process is as follows: the system first obtains the current time, and determines the current time stage according to the preset time stage division. For example, assuming that the current time is 10:30, according to the system settings, it is in the daytime stage (8:00-18:00). Next, the system reads the preset weight adjustment rules for the daytime stage from the configuration file. Assuming that the rule is set to the daytime stage, the initial attention coefficient needs to be multiplied by the weight coefficient 0.9. Then, the system calls the preliminary attention coefficient obtained in step S402, assuming that the coefficient is 0.7. According to the adjustment rules, the adjusted attention coefficient is calculated: 0.7×0.9=0.63. In this way, the system adjusts the preliminary attention coefficient according to different time stages, so that the attention coefficient can more accurately reflect the degree of attention of medical staff to patient infusion information under different time conditions, and provide a basis for obtaining a more accurate overall effective score in the future.

[0127] Step S404, adjusting the attention coefficient again based on the preset condition weight adjustment method and the patient's condition included in the patient's condition related information.

[0128] Preset disease state weight adjustment method: a set of pre-established rules and methods for adjusting the attention coefficient according to the patient's disease state. Different disease states mean that the patient has different needs for infusion monitoring. This rule is used to reasonably adjust the attention coefficient to make it more in line with the patient's actual needs.

[0129] Patient's condition status: a description of the severity and development trend of the patient's current condition, such as stable condition, worsening condition, sudden symptoms, etc. These conditions are determined by the patient's various physiological indicators, disease diagnosis results, clinical judgment of medical staff and other information.

[0130] The general process is as follows: the system first obtains the patient's relevant medical data from the hospital information management system, and judges the patient's condition through the preset condition assessment model. Assume that based on the acquired data and model analysis, a patient's condition is judged to be "worsened." Next, the system reads the preset condition weight adjustment method. If the method stipulates that when the patient's condition is "worsened," the attention coefficient must be multiplied by the weight coefficient 1.5. At this time, the system calls the attention coefficient obtained in step S403, assuming it is 0.63. Calculate the adjusted attention coefficient according to the adjustment method: 0.63×1.5=0.945.

[0131] Step S405, according to the preset multi-dimensional data influence rules, the fluctuation of vital signs within the preset time range contained in the patient's condition-related information and the frequency of infusion speed changes contained in the infusion task-related information are comprehensively considered to make a final correction to the attention coefficient.

[0132] Preset multi-dimensional data impact rules: A set of pre-set rules used to comprehensively consider the impact of multiple data dimensions (such as fluctuations in vital signs, frequency of changes in infusion speed, etc.) on the attention coefficient, so as to make final corrections to the attention coefficient so that it can more accurately reflect the urgency of the patient's infusion situation and the degree of attention that medical staff should give.

[0133] Frequency of infusion speed change: The number of times the patient's infusion speed changes within a certain period of time reflects the stability of the infusion process. For example, if the infusion speed has been adjusted 3 times in the past 2 hours, the frequency of infusion speed change is 1.5 times per hour. Fluctuation of vital signs within a preset time range: refers to the fluctuation and changes of vital signs such as heart rate, blood pressure, body temperature, etc. within a preset time range (such as the past 12 hours). A larger fluctuation may mean that the patient's condition is unstable and the need for infusion monitoring is higher.

[0134] The general process is described as follows: the system first obtains the frequency data of infusion speed changes from the infusion equipment management system. Assuming that the infusion speed of a patient has changed 4 times in the past 3 hours, the frequency of infusion speed change is about 1.33 times per hour. At the same time, the vital signs fluctuation data of the patient within the preset time range is obtained from the hospital information management system. For example, in the past 12 hours, the patient's heart rate fluctuates between 60-100 times / minute, with a large fluctuation range. Then, the system reads the preset multi-dimensional data impact rules. If the rules stipulate that when the frequency of infusion speed change is greater than once per hour and the fluctuation range of vital signs exceeds a certain threshold (assuming that the heart rate fluctuation range is set to exceed 20 times / minute), the attention coefficient needs to be multiplied by the correction coefficient 1.2. At this time, the system calls the attention coefficient obtained in step S404, assuming it is 0.945. According to the rules, the final corrected attention coefficient 0.945×1.2=1.134 is calculated.

[0135] Step S406: multiply the determined attention coefficient by the overall score of each patient to obtain a final overall effective score.

[0136] Overall effective score: The result obtained by multiplying the determined attention coefficient with the overall score of each patient. This score comprehensively considers the medical staff's attention to the patient's infusion information and the patient's infusion situation's assessment. It can more comprehensively reflect the priority and importance of the patient's infusion situation in actual medical scenarios, and provide a more accurate basis for the subsequent sorting of patient infusion data.

[0137] Step S407, automatically sorting the patient infusion data from high to low according to the final overall effective score, and presenting the sorting result on the display interface of the mobile terminal held by the medical staff.

[0138] Automated sorting: The process of automatically sorting patient infusion data based on the final overall effective score using computer programs and algorithms. The purpose of sorting is to provide medical staff with a clear priority list of patient infusion situations, so that they can quickly understand which patients need to be focused on. Mobile terminal display interface: The screen area on the mobile devices (such as smartphones and tablets) used by medical staff to display the sorting results of patient infusion data. This interface presents the sorted patient infusion information in an intuitive way, helping medical staff to obtain key information in a timely manner.

[0139] An intelligent infusion assessment reminder method also includes a voice interaction step, which is as follows: Step Sd00, analyzing whether the infusion pump receives the patient's voice description information about the infusion feeling during the infusion, and defining that the infusion pump is provided with the function of receiving the voice description information and transmitting it. If yes, execute step Se00; if no, execute step Sf00.

[0140] Among them, the infusion pump is a medical device used to accurately control the infusion speed and volume. In this intelligent infusion assessment reminder method, the infusion pump is given the additional function of receiving and transmitting the patient's voice expression information. Voice expression information: the voice data generated by the patient expressing his or her infusion experience during the infusion, such as "I feel a little pain in my arm" and "The infusion speed is too fast".

[0141] Step Se00, transmit the patient's voice description information about the infusion feeling during the infusion to the mobile terminal held by the medical staff, receive the voice description information fed back by the medical staff, transmit it back to the infusion pump and make a voice description of the medical staff.

[0142] Step Sf00, maintain the original settings.

[0143] Among them, the original settings refer to the current operating status settings of the intelligent infusion assessment reminder system when it has not received the patient's voice expression information about the infusion experience, including the infusion parameter settings of the infusion pump, the communication configuration of the system with other devices or systems, etc.

[0144] An intelligent infusion assessment reminder method further includes the following steps after analyzing whether the infusion pump receives the patient's voice expression information about the infusion feeling during the infusion: Step Sd01, if it is confirmed that the infusion pump has received the patient's valid voice expression information, the patient's voice expression information is input into the pre-built bidirectional long short-term memory network keyword matching model, and the preset emergency keyword matching result is output. The keyword urgency is graded to obtain the keyword matching score K.

[0145] Patient's effective voice expression information: The patient's voice expression about the infusion experience that is received and confirmed by the infusion pump, has practical significance and can be used for subsequent analysis. For example, the patient clearly expresses "pain at the infusion site" instead of meaningless noise or words unrelated to the infusion experience.

[0146] Bidirectional long short-term memory network keyword matching model: A model based on deep learning, which is specially used to analyze the input text information (here, the text converted from the patient's voice expression). By learning the language patterns and semantic features in a large amount of text data, it can identify the part of the text that matches the preset emergency keywords and output the matching results. This model can capture the information of the text context and effectively process sequence data. Preset emergency keywords: pre-set words or phrases related to the patient's infusion emergency, such as "unbearable pain", "palpitation", "difficulty breathing", etc. These keywords are determined based on medical experience and common adverse reactions to infusion, and are used to measure the urgency of the patient's voice expression.

[0147] Keyword urgency grading: The preset emergency keywords are graded according to their urgency. For example, extremely serious conditions such as "cardiac arrest" are graded as the highest level, and "mild pain" is graded as a lower level. Grading can more intuitively quantify the keyword matching results.

[0148] Keyword matching score K: A quantitative score calculated based on the output results of the bidirectional long short-term memory network keyword matching model and the keyword urgency grading. It is used to indicate the degree of correlation between the patient's voice expression information and the emergency situation. The higher the score, the higher the urgency may be.

[0149] The general process is as follows: the system first obtains the patient's effective voice expression information obtained in step Sd00 from the temporary data storage module. Assume that the information obtained is "I feel panic and my heart beats very fast". Then, the information is input into the loaded bidirectional long short-term memory network keyword matching model. The model analyzes the input text and identifies the semantic features and language patterns in the text through its internal neural network structure. For example, the model may split the input text into words or character sequences and match them with the patterns learned by itself. In this example, the model recognizes the preset emergency keyword "panic". Then, the system queries the urgency level corresponding to "panic" based on the preset emergency keywords and keyword urgency classification. Assume that "panic" is classified as medium urgency, corresponding to a specific basic score (such as 5 points). The system then adjusts the basic score according to the matching degree of the model (such as the accuracy of the matching, whether there is fuzzy matching, etc.), and finally calculates the keyword matching score K. For example, due to the accuracy of this matching, 1 point is added to the basic score of 5 points, and the keyword matching score K is 6 points.

[0150] Step Sd02, input the patient's speech expression information into the medical speech sentiment analysis model based on convolutional neural network, and output the speech sentiment intensity score S.

[0151] Medical speech sentiment analysis model based on convolutional neural network: A model built with the help of convolutional neural network technology, which is specifically used to analyze the emotional intensity of patient speech in medical scenarios. By constructing structures such as convolutional layers and pooling layers, convolutional neural networks can automatically extract features from speech signals to determine the intensity of emotions expressed by speech, such as anxiety, pain, etc.

[0152] Speech emotion intensity score S: The model outputs a quantitative result after analyzing the input patient speech information, which is used to measure the intensity of the emotion expressed by the patient in the speech. The higher the score, the higher the intensity of the patient's emotion, such as extreme pain or great anxiety; the lower the score, the relatively calm or mild emotion.

[0153] Step Sd03, use the UWB positioning system to obtain the patient's real-time location coordinates, combine the hospital map to determine the area, use the hospital regional location assessment model based on hierarchical analysis method, input the emergency treatment resources of each area of ​​the hospital and the frequency of historical emergency events, output the regional priority weight, and calculate the location association score P.

[0154] UWB positioning system: Ultra-Wideband positioning system, which can accurately locate the target by sending and receiving nanosecond non-sinusoidal narrow pulses. In the intelligent infusion scenario, it is used to obtain the patient's position coordinates in real time. Its positioning accuracy is high and can meet the medical scenario's demand for accurate positioning of the patient's position. Real-time location coordinates: The patient's specific location information at a certain moment in the hospital space, expressed in the form of coordinate values, such as (x, y) in a two-dimensional coordinate system and (x, y, z) in a three-dimensional coordinate system. The coordinates reflect the patient's specific location in the hospital and help determine the area where he is located. Hospital map: A digital representation of the internal spatial layout of a hospital, including the division of various areas (such as ward areas, operating room areas, nurse stations, etc.), channel locations, facility distribution, etc. It is an important basis for associating the patient's location coordinates with the actual area of ​​the hospital. Hospital regional location assessment model based on analytic hierarchy process: A model built using analytic hierarchy process (AHP) to evaluate the priority of different areas of a hospital in responding to emergencies. The analytic hierarchy process decomposes complex problems into multiple levels, establishes a judgment matrix, and calculates the relative weights of elements at each level, thereby comprehensively evaluating the emergency handling capabilities of different areas. The model takes into account factors such as the emergency handling resources of each area of ​​the hospital (such as the number of emergency equipment, medical staff, etc.), the frequency of historical emergency events, and outputs regional priority weights.

[0155] Regional priority weight: A value calculated based on the hospital regional location assessment model based on the analytic hierarchy process, reflecting the relative importance or priority of each area of ​​the hospital in responding to emergencies. The higher the weight, the more abundant the resources in the area for emergency treatment or the higher the frequency of historical emergencies, and the more attention it needs. Location correlation score P: A quantitative score calculated by combining the patient's real-time location coordinates, the area where he is located, and the area priority weight. It is used to measure the degree of correlation between the patient's location and emergency treatment. The higher the score, the higher the emergency treatment priority of the location.

[0156] Step Sd04, extracting the infusion start time from the infusion pump and the associated system, analyzing and obtaining the infusion duration, inputting the infusion duration into the disease deterioration probability prediction model based on time series analysis, and outputting the adjusted time weight coefficient α.

[0157] Infusion start time: The specific moment when the patient starts infusion treatment, accurate to minutes or even seconds, is an important basic data for calculating the infusion duration and predicting the probability of subsequent worsening of the disease. Infusion duration: The time interval from the start of infusion to the current moment, used to measure the duration of the patient's infusion. This duration is of great significance in analyzing the trend of the patient's condition and assessing the possibility of worsening of the condition. Deterioration probability prediction model based on time series analysis: A model built using time series analysis methods, specifically used to predict the probability of a patient's condition deteriorating based on time-related data such as infusion duration. Time series analysis uses statistics and modeling of historical data to mine the laws and trends of data changes over time, thereby predicting future situations. In this model, time series data such as infusion duration are input, and the corresponding probability of deterioration is output, which is further converted into an adjusted time weight coefficient for subsequent urgency assessment. Adjusted time weight coefficient α: a value obtained through certain conversion or calculation based on the output of the disease deterioration probability prediction model based on time series analysis. This coefficient reflects the influence weight of the correlation between the duration of infusion and the possibility of disease deterioration on the urgency assessment. The larger the coefficient, the higher the proportion of the infusion duration factor in the urgency assessment.

[0158] Step Sd05, using the emotion weight dynamic adjustment model based on machine learning, input the historical data of the patient's voice emotion during the past infusion and the current severity of the disease, and output the dynamically determined emotion intensity weight coefficient β.

[0159] Dynamic adjustment model of emotion weight based on machine learning: A model built using machine learning algorithms that can dynamically determine the emotion intensity weight coefficient based on input data. It adapts to the changes in the weight of emotion intensity in the urgency assessment of different patients and different stages of the disease by learning patterns and rules from a large amount of historical data. Historical data of the patient's voice emotion during the past infusion: Record the emotional characteristics and patterns of the patient's voice during the previous infusion. For example, some patients are used to expressing discomfort with the infusion in a stronger tone, while some patients are relatively mild. These long-term expression habits and emotional characteristics constitute the historical data of the patient's voice emotion during the past infusion. Current severity of illness: reflects the severity of the patient's physical condition at the moment, and is determined by medical staff based on the patient's symptoms, test results, vital signs and other information, such as mild, moderately severe, critical, etc. Emotional intensity weight coefficient β: A value used to measure the proportion of voice emotion intensity in the urgency assessment. The model output is dynamically adjusted through the emotion weight based on machine learning. The coefficient changes dynamically based on the patient's historical data of voice emotion during past infusions and the current severity of the disease, so as to more accurately reflect the impact of emotional factors on the urgency.

[0160] Step Sd06, adopting a location weight determination model based on the hierarchical analysis method, inputting the hospital area emergency treatment priority and patient location information, and outputting a reasonable location weight coefficient γ.

[0161] Among them, the location weight determination model based on the analytic hierarchy process is a model for determining the location weight coefficient using the analytic hierarchy process (AHP). The analytic hierarchy process decomposes complex problems into multiple levels, and comprehensively evaluates the impact of different factors on the target by constructing a judgment matrix and calculating the relative weights of the elements at each level. In this model, it is used to comprehensively consider the emergency treatment priority of the hospital area and the patient location information to determine a reasonable location weight coefficient. Hospital regional emergency response priority: Based on the emergency response resources of each area of ​​the hospital (such as the number and sophistication of emergency equipment, the number and qualifications of professional emergency medical staff, etc.) and the frequency of historical emergency events, the priority of each area in emergency response is determined through a certain evaluation method. Areas with higher priorities have stronger resource guarantees and higher urgency when responding to emergencies. Patient location information: The real-time location coordinates of patients in the hospital are obtained through positioning technologies such as UWB positioning systems. Combined with the hospital map, the specific area where the patient is located can be determined, such as the ward area, corridor, near the nurse station, etc. Location weight coefficient γ: A value calculated by the location weight determination model based on the hierarchical analysis method, which is used to measure the proportion of the patient location factor in the urgency assessment. This coefficient comprehensively considers the emergency treatment priority of the hospital area and the patient location information. The larger the coefficient, the greater the impact of the patient location factor on the urgency assessment.

[0162] Step Sd07, substituting the calculated α, K, β, S, γ, and P into a preset emergency evaluation value calculation formula to obtain an emergency evaluation value, wherein the preset emergency evaluation value calculation formula is as follows: =α×K+β×S+γ×P.

[0163] Step Sd08: The patient's voice information and the emergency assessment value are recorded. , real-time location information and infusion-related parameters are packaged and transmitted to the medical staff’s mobile terminal.

[0164] The patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Step S40a, within a preset time period, the number of times the medical staff and the medical staff give voice instructions or voice descriptions of information related to different patients' infusions is recorded through the voice interaction system of the mobile terminal. , and simultaneously uses speech recognition and semantic analysis technology to identify the urgency-related keywords mentioned in each speech and mark the relevant speech.

[0165] Among them, the preset time period range is a pre-set time period used to limit the time span of data statistics, such as the past 24 hours, a week, etc., so as to collect voice interaction data related to patient infusion within that specific time.

[0166] Voice interaction system for mobile terminals: a software system integrated on mobile devices used by medical staff (such as smartphones and tablets), which has functions such as voice command reception, voice expression information collection, voice recognition and semantic analysis, making it convenient for medical staff to interact with the system through voice.

[0167] Voice commands or voice description information: During work, medical staff send voice commands to mobile terminals to query patient infusion-related information, such as "check Zhang San's infusion speed", or descriptions of the patient's infusion status, such as "Li Si's infusion reaction is not very good."

[0168] Speech recognition and semantic analysis technology: Speech recognition technology converts the voice signals input by medical staff into text form that can be understood by computers; semantic analysis technology understands the converted text and extracts key information, such as identifying the patient's name mentioned in the text, infusion-related parameters, and keywords related to the degree of urgency.

[0169] Urgency-related keywords: pre-defined words or phrases that are closely related to the patient's infusion emergency, such as "emergency", "severe reaction", "abnormal heartbeat", etc. These keywords can help the system quickly determine the urgency of the content involved in the voice.

[0170] Mark relevant speech: After the system recognizes that the speech contains keywords related to urgency, it will specially mark the speech for subsequent targeted statistics and analysis.

[0171] Step S40b: Count the frequency of occurrence of emergency keywords in each patient's related speech .

[0172] Occurrence frequency: The ratio of the number of times a specific emergency keyword appears in all speech related to a patient to the total number of speech related to the patient, which is used to measure how frequently emergency situations are mentioned in speech related to the patient. A higher frequency may mean that the patient's infusion situation needs more attention.

[0173] Step S40c, using a preset comprehensive concern factor calculation formula to calculate the comprehensive concern factor of each patient, the comprehensive concern factor calculation formula is as follows: ; in, is the comprehensive attention factor, The number of voice commands The weight coefficient of The frequency of urgent keywords The weight coefficient of Assess the urgency level The weight coefficient of .

[0174] Comprehensive attention factor: A value that comprehensively considers multiple factors such as the number of voice commands, the frequency of emergency keywords, and the urgency assessment value, used to quantify the degree of attention that medical staff pay to the patient's infusion situation. The higher the factor, the more attention the patient's infusion situation deserves.

[0175] Step S40d, multiplying each patient's comprehensive concern factor by the patient's original overall score to obtain an overall effective score.

[0176] Step S40e, automatically sorting the patient infusion data from high to low according to the overall effective score, and presenting the sorting result on the display interface of the mobile terminal held by the medical staff.

[0177] Furthermore, the intelligent infusion assessment reminder method also includes a processing step in which the intelligent infusion pump and the intelligent bed work in coordination, which is as follows: The smart infusion pump undertakes the key task of real-time monitoring of the infusion status. It continuously collects important information such as infusion progress, remaining time and infusion speed at intervals of 1 minute. Once it detects that the remaining infusion time is less than 30 minutes, it immediately sends a coordination signal to the smart bed. The smart bed works synchronously through pressure sensors and angle sensors. The pressure sensor collects pressure distribution data of various parts of the patient's body once a second, and the angle sensor collects the position angle data of the bed every 5 minutes. These data are transmitted to the smart infusion evaluation reminder system in real time, providing comprehensive and timely data support for subsequent analysis and decision-making.

[0178] After receiving the pressure data of the smart bed, the system first uses the moving average filtering algorithm to preprocess the data, effectively remove the noise in the data and improve the data quality. Subsequently, the system calculates the mean and standard deviation of the processed pressure data to determine the normal pressure range as the mean plus or minus 2 times the standard deviation. When the pressure value exceeds the normal range for 3 consecutive times, the system determines that the pressure is abnormal and divides the abnormal level according to the degree of excess, which is divided into mild (exceeding the normal range by 1-1.5 times the standard deviation), moderate (exceeding 1.5-2 times the standard deviation) and severe (exceeding 2 times the standard deviation or more), providing a basis for subsequent precise adjustments.

[0179] According to the level of pressure abnormality, the system adjusts the patient's overall score accordingly. If the pressure distribution is normal, the overall score remains unchanged; if it is a mild abnormality, the overall score is reduced by 3 points; moderate abnormality is reduced by 6 points; severe abnormality is reduced by 9 points. This adjustment mechanism closely combines the pressure abnormality with the patient's overall assessment, so that the overall score can more accurately reflect the patient's infusion status and provide more valuable reference information for medical staff.

[0180] The system re-evaluates the reminder level based on the adjusted overall score, which is divided into high emergency (score ≥ 80 points), medium emergency (60-79 points), and low emergency (<60 points). If the reminder level changes, the system immediately sends a reminder message to the medical staff's mobile terminal to inform the level change and the reason. At the same time, the collaborative working parameters of the infusion pump and the bed are adjusted according to the reminder level and the degree of pressure abnormality. If the reminder level increases and the pressure is severely abnormal, the infusion speed is reduced by 30% and the head of the bed is raised by 20°; if the reminder level decreases and the pressure improves, the infusion speed is increased by 10%, and the bed position gradually returns to the initial state to ensure the safety and comfort of the infusion process.

[0181] The smart infusion pump and smart bed continuously collect data and feed it back to the system in real time. The system updates the overall score, reminder level and collaborative working status in real time based on the new data. If the pressure continues to be abnormal, the system will adjust the collaborative parameters again or suspend the infusion and remind the medical staff; if the pressure improves, the infusion speed and bed position will be gradually restored. In addition, the system regularly analyzes the data in the collaborative process, summarizes the correlation between abnormal pressure and infusion conditions, and continuously optimizes the score adjustment rules and collaborative working parameter adjustment algorithms to improve the accuracy and effectiveness of the collaborative work of the smart infusion pump and the smart bed.

[0182] Based on the same inventive concept, an embodiment of the present invention provides an intelligent infusion assessment reminder system, including a memory and a processor, wherein the memory stores a program that can be executed on the processor to implement the following Figure 1 to Figure 2 Procedure for either method.

[0183] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent infusion assessment reminder method, characterized in that: include: Obtain information related to the patient's condition and infusion task entered into the system; Determine the score of the corresponding indicator according to the mapping relationship between the range of each indicator data in the disease-related information and the infusion task-related information and the score; According to the preset time decay formula, the weight of each indicator score changing over time is calculated according to the time elapsed since the indicator data was collected. At the same time, the gene polymorphism data and past medical history are extracted from the patient's condition-related information and input into the pre-built individual difference correction coefficient analysis model to output the correction coefficient for each indicator score; Combine the original scores of each indicator in the patient's condition-related information and the infusion task-related information, multiply them by the corresponding time decay weight and the individual difference correction coefficient, and then add them up to calculate the overall score of each patient. Automatically sort the patient's infusion data in descending order according to the overall score, and display them on the display interface of the mobile terminal held by the medical staff; According to the mapping relationship between the interval range of the overall score and the reminder level, the reminder level is analyzed and determined, wherein the reminder level is divided into three levels: high urgency, medium urgency, and low urgency; Obtain the ambient noise value and working status of the environment where the medical staff is currently working, perform fuzzy processing on the acquired ambient noise value and working status of the medical staff, convert them into elements in the fuzzy set, and establish the logical relationship between the ambient noise, working status of the medical staff and the reminder level based on the pre-established fuzzy rule base; Through fuzzy reasoning, corresponding fuzzy output results are generated for different reminder levels. Then, through defuzzification calculation, the specific reminder method corresponding to the reminder level is determined, and the corresponding specific reminder method is executed.

2. The intelligent infusion assessment reminder method according to claim 1, characterized in that: After executing the corresponding specific reminder method, it also includes: Continuously analyze whether there are high-urgency alerts; If not, maintain the original setting; If yes, then obtain the scene information of the patient, and collect the working status, location information, workload of the medical staff, and the severity and urgency of the patient's condition in real time. The scene information includes the general ward scene, emergency rescue scene, operating room scene and emergency room scene; According to the preset quantitative scoring method, the working status, location information, workload, environmental noise value of medical staff, and the severity and urgency of the patient's condition are quantitatively scored; Based on the quantitative scoring of medical staff's work status score, location information score, workload score, and environmental noise value score, the preset optimal task allocation method based on particle swarm and ant colony algorithm is used to analyze, determine, and execute the optimal task allocation plan.

3. The intelligent infusion assessment reminder method according to claim 2, characterized in that: The optimal task allocation methods based on particle swarm and ant colony algorithms include: A preset number N of ants is randomly generated, each ant corresponds to a medical staff allocation plan, a pheromone matrix with the number of rows equal to the number of medical staff M and the number of columns equal to 1 is created, the initial value is set to a preset constant T0, and a preset number P of particles is synchronously generated, each particle represents a task allocation plan, and the initial position and speed parameters are preset; The work status score, location information score, workload score, environmental noise value score of the medical staff obtained by quantitative scoring, as well as the severity score and urgency score of the patient's condition are input into the preset fitness calculation function. According to the pre-set weight adjustment rules for different scenarios of the patient, the corresponding weight is assigned to each scoring factor. The fitness calculation function multiplies each score with the corresponding weight and accumulates them to obtain the fitness value of the task allocation scheme represented by each ant and each particle; Particles update their positions and speeds according to the preset rules and the optimal positions of themselves and the particle group. Each iteration uses a preset function to calculate the fitness value. If the new value is better than the historical optimal value, the corresponding extreme value is updated. After the particle swarm finds the global optimal solution, it increases the pheromone concentration of the ant colony's high-quality paths according to the preset rules. When the ants choose a path, they incorporate the particle swarm's global optimal factors into the selection probability calculation according to specific weights. When the particles are updated, if the pheromone concentration in a certain area of ​​the ant colony exceeds the preset threshold, the probability of moving to that area is increased according to the preset rules. Ants select medical staff according to the preset path selection probability formula. The path selection probability formula integrates pheromone concentration, heuristic information and particle swarm global optimal factors. In different scenarios, the weights of each factor are dynamically adjusted according to the preset weight adjustment formula to achieve path selection. Each time the ants complete the path selection, the pheromone matrix will be updated according to the preset pheromone update formula. When the preset number of iterations T is reached or the fitness improvement in n consecutive iterations is less than the preset value, the ant colony iteration is stopped; According to the preset solution screening formula, the solutions represented by ants and particles are evaluated, and the solution with the highest fitness is selected; Assign high-emergency patient tasks to the medical staff corresponding to the plan, and send the patient's condition and scenario to the mobile terminal held by the medical staff through the hospital information channel according to the preset information push rules.

4. The intelligent infusion assessment reminder method according to claim 1, characterized in that: After executing the corresponding specific reminder method, it also includes: Analyze whether the mobile terminal receives voice instructions from medical staff to view infusion-related information of the patient in charge; If yes, retrieve the patient's infusion related information according to the voice command and make a voice reply; If not, keep the original setting.

5. The intelligent infusion assessment reminder method according to claim 4, characterized in that: The patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Within a preset time period, the voice commands received by the mobile terminal for viewing the infusion-related information of different patients are recorded, and the timestamp corresponding to each command is marked; According to the number of voice commands, the attention coefficient is preliminarily determined according to the corresponding relationship between the preset number range and the attention coefficient; According to the preset time stage weight adjustment rules, the initial attention coefficient is adjusted according to different time stages; According to the preset condition status weight adjustment method, the attention coefficient is adjusted again in combination with the patient's condition status included in the patient's condition related information; According to the preset multi-dimensional data influence rules, the fluctuation of vital signs within the preset time range contained in the patient's condition-related information and the frequency of changes in the infusion speed contained in the infusion task-related information are comprehensively considered to make a final correction to the attention coefficient; The determined attention coefficient was multiplied by each patient's overall score to obtain the final overall effective score; According to the final overall effective score, the patient infusion data is automatically sorted from high to low, and the sorting results are presented on the display interface of the mobile terminal held by the medical staff.

6. The intelligent infusion assessment reminder method according to claim 1, characterized in that: It also includes voice interaction steps, as follows: Analyze whether the infusion pump receives the patient's voice description information about the infusion feeling during the infusion, and define that the infusion pump is provided with the function of receiving the voice description information and transmitting it; If yes, the patient's voice description of the infusion feeling during the infusion is transmitted to the mobile terminal held by the medical staff, and the voice description information fed back by the medical staff is received and transmitted back to the infusion pump and the medical staff's voice description is made; If not, keep the original setting.

7. The intelligent infusion assessment reminder method according to claim 6, characterized in that: After analyzing whether the infusion pump receives the patient's voice description information about the infusion feeling during the infusion, the following steps are also included: If it is confirmed that the infusion pump has received the patient's valid voice expression information, the patient's voice expression information is input into the pre-built bidirectional long short-term memory network keyword matching model, and the preset emergency keyword matching result is output, and the keyword urgency is graded to obtain the keyword matching score K; The patient's speech expression information is input into the medical speech sentiment analysis model based on convolutional neural network, and the speech sentiment intensity score S is output; The real-time location coordinates of the patient are obtained by using the UWB positioning system, and the area is determined by combining the hospital map. The hospital regional location evaluation model based on the hierarchical analysis method is used to input the emergency treatment resources of each area of ​​the hospital and the frequency of historical emergency events, output the regional priority weight, and calculate the location association score P; Extract the infusion start time from the infusion pump and related systems, analyze and obtain the infusion duration, input the infusion duration into the disease deterioration probability prediction model based on time series analysis, and output the adjusted time weight coefficient α; Using the emotion weight dynamic adjustment model based on machine learning, the historical data of the patient's voice emotion during the past infusion and the current severity of the disease are input, and the dynamically determined emotion intensity weight coefficient β is output; A location weight determination model based on the hierarchical analysis method is used to input the hospital area emergency treatment priority and patient location information, and output a reasonable location weight coefficient γ; Substitute the calculated α, K, β, S, γ, and P into the preset emergency assessment value calculation formula to obtain the emergency assessment value, where the preset emergency assessment value calculation formula is as follows: =α×K+β×S+γ×P; The patient's voice-expressed information is combined with an urgency assessment value , real-time location information and infusion-related parameters are packaged and transmitted to the medical staff’s mobile terminal.

8. The intelligent infusion assessment reminder method according to claim 7, characterized in that: The patient's infusion data is automatically sorted in descending order according to the overall score, and displayed on the display interface of the mobile terminal held by the medical staff, including: Within a preset time period, the number of voice instructions or voice descriptions of medical staff and feedback to medical staff on infusion-related information for different patients is recorded through the voice interaction system of the mobile terminal. , and simultaneously use speech recognition and semantic analysis technology to identify the urgency-related keywords mentioned in each speech and mark the relevant speech; Count the frequency of emergency keywords in each patient's relevant speech ; The comprehensive attention factor of each patient is calculated using the preset comprehensive attention factor calculation formula. The comprehensive attention factor calculation formula is as follows: ； in, is the comprehensive attention factor, The number of voice commands The weight coefficient of The frequency of urgent keywords The weight coefficient of Assess the urgency level The weight coefficient of Multiply each patient's comprehensive concern factor by the patient's original overall score to obtain the overall effective score; According to the overall effective score, the patient infusion data is automatically sorted from high to low, and the sorting results are presented on the display interface of the mobile terminal held by the medical staff.

9. An intelligent infusion assessment reminder system, characterized in that: It comprises a memory, a processor and a program stored in the memory and executable on the processor, wherein the program can be loaded and executed by the processor to implement an intelligent infusion assessment reminder method as described in any one of claims 1 to 8.

Images