Medical monitoring method based on multi-mode prompt interaction

Abstract

The invention discloses a medical monitoring method based on multi-modal prompt interaction, which comprises the following steps: acquiring vital sign data of a patient, performing scoring and sorting according to patient information and the vital sign data to obtain patient monitoring scores, and forming a patient monitoring sequence; obtaining the service capability data, the current workload and the position data of the medical personnel, carrying out the sorting and scoring according to the service capability data, the current workload and the position data of the medical personnel, obtaining the state score of the medical personnel, and forming a medical personnel distribution sequence; according to the priority order of the patients in the patient monitoring queue, selecting the optimal medical staff in the medical staff distribution queue; and according to the patient monitoring score, generating a multi-modal prompting strategy, and according to the multi-modal prompting strategy, carrying out multi-modal prompting on the selected optimal medical personnel.

Description

Technical Field

[0001] The present invention relates to the technical field of medical monitoring, and particularly to a medical monitoring method based on multi-modal prompt interaction. Background Art

[0002] In the inpatient departments and intensive care units (ICUs) of modern hospitals, medical staff need to monitor the vital signs (such as heart rate, blood pressure, blood oxygen saturation, and respiratory rate, etc.) of multiple patients simultaneously to ensure the safety of patients.

[0003] However, with the popularization of medical devices and the increase in the number of patients, the corresponding medical staff resources are significantly insufficient. Although the existing technology can remotely prompt and notify medical staff through a centralized medical care system, the current medical monitoring system mainly relies on central vision (such as monitor display) and auditory (such as alarm sound) prompts. This single prompt method is prone to information overload. Especially when multiple patients issue alarms simultaneously, the vision and hearing of medical staff are occupied. Especially when medical staff are performing other tasks (such as emergency care, equipment operation), they may not be able to notice the alarm information in time. Since the existing system mainly relies on vision and hearing, the cognitive load of medical staff will increase significantly, further reducing the response speed and accuracy to alarm information.

[0004] In addition, in the scenario of multiple patients and multiple tasks, the existing system usually cannot effectively distinguish the priority of alarms, nor does it have a hierarchical prompt strategy for the importance of different patients and the urgency of alarms. This may lead to medical staff being unable to obtain key alarm information in time. Further, when the existing system processes alarm information, it usually adopts a simple first-come-first-served or fixed rule, and fails to fully consider the priority of alarm information and the importance of patients. This may lead to unreasonable resource allocation and delay in the treatment of high-priority patients.

[0005] Designing a medical monitoring method based on multi-modal prompt interaction for the above problems existing in the prior art is the purpose of the research of the present invention. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to propose a medical monitoring method based on multi-modal prompt interaction, which can solve the above problems.

[0007] The present invention provides a medical monitoring method based on multi-modal prompt interaction, including: Obtaining the vital sign data of patients, scoring and sorting according to patient information and vital sign data to obtain a patient monitoring score, and forming a patient monitoring sequence; Obtain the professional ability data, current workload, and location data of medical staff, sort and score according to the professional ability data, current workload, and location data of medical staff to obtain the medical staff status score, and form a medical staff allocation sequence; Select the optimal medical staff from the medical staff allocation queue according to the priority order of patients in the patient monitoring queue; Generate a multimodal prompt strategy based on the patient monitoring score, and perform multimodal prompts on the selected optimal medical staff according to the multimodal prompt strategy.

[0008] Furthermore, the obtaining of the vital sign data of the patient, scoring and sorting according to the patient information and vital sign data to obtain the patient monitoring score, and forming the patient monitoring sequence includes: Obtain the real-time heart rate data, blood pressure data, blood oxygen saturation data, respiratory rate data, and body temperature of the patient; Through the real-time heart rate data of the patient 、blood pressure data 、blood oxygen saturation data 、respiratory rate data 、body temperature Calculate the physical sign score of each index of the patient ; Calculate the monitoring score of the patient through the physical sign score of each index of the patient , and the specific calculation formula is as follows: , where, is the weight of the corresponding index; Obtain the patient's past medical history data, and mark the key monitoring vital sign data through the patient's past medical history data. The specific calculation formula is as follows: , where, is the medical history correction item, is the global weight coefficient affected by the medical history.

[0009] Furthermore, the calculating of the physical sign score of each index of the patient through the real-time heart rate data 、blood pressure data 、blood oxygen saturation data 、respiratory rate data 、body temperature includes: The physical sign score formula for heart rate data is ; ; The physical sign score formula for blood pressure data is ; The physical sign score formula for blood oxygen saturation data is ; The physical sign scoring formula for respiratory rate data is ; The physical sign scoring formula for body temperature data is .

[0010] Furthermore, calculating the patient's monitoring score through the physical sign scores of each index of the patient includes: If three consecutive index measurements satisfy |Δv / Δt| ≥ 10%, where Δv is the change in the vital sign value and Δt is the time interval, then the weight is adjusted through the weight adjustment formula, and the weight adjustment formula is as follows: , where is the width of the normal range of each index.

[0011] Furthermore, the calculation formula of the medical history correction item is as follows: , where is the enhancement coefficient of the medical history type k for the relevant physical sign , is the physical sign score associated with the medical history k, is the baseline risk addition of the medical history k, and n is the total number of medical history types the patient has.

[0012] Furthermore, obtaining the professional ability data, current workload, and location data of medical staff, sorting and scoring according to the professional ability data, current workload, and location data of medical staff, and obtaining the medical staff status score to form the medical staff allocation sequence includes: Calculating the fitness A between the professional ability data of medical staff and the patient's medical record data through the semantic similarity model; Calculating the current workload U of medical staff according to the current task quantity and working hours of medical staff; Calculating the path cost C from the current location to the patient location through the moving distance d of medical staff and the required equipment preparation time , and the calculation formula is as follows: , where , are the weights corresponding to the moving distance d and the required equipment preparation time respectively; Calculating the medical staff status score according to the data fitness A, the previous workload U, and the path cost C, and the calculation formula is as follows: ; The S205 medical staff status scores are sorted from high to low to form a medical staff allocation sequence.

[0013] Furthermore, the calculation of the current workload U of the medical staff according to the current task quantity and working hours of the medical staff includes: Obtain the current task data volume current_tasks of the medical staff and its maximum bearable task quantity max_capacity; Obtain the continuous working time t, and calculate the current workload U through the current task data volume current_tasks of the medical staff, its maximum bearable task quantity max_capacity, and the continuous working time t. The calculation formula is as follows: .

[0014] Furthermore, the selection of the optimal medical staff from the medical staff allocation queue according to the priority order of the patients in the patient monitoring queue includes: Select a number of patients with the highest priority from the patient monitoring sequence to form a sequence of patients to be allocated; For each patient in the sequence of patients to be allocated , traverse all medical staff in the medical staff allocation sequence , calculate the score of each patient-medical staff combination, and correct the score of each patient-medical staff combination through the distance , and the calculation formula is as follows: , Wherein, is the monitoring score of patient , is the status score of medical staff , is the physical distance between patient and medical staff , is the distance attenuation coefficient, is the distance correction factor; Form a combination score matrix with the scores of each patient-medical staff combination, and select the patient-medical staff combination score that maximizes the global total score to obtain the allocation form.

[0015] Furthermore, the generation of a multimodal prompt strategy according to the patient monitoring score and the multimodal prompting of the selected optimal medical staff according to the multimodal prompt strategy include: According to the allocation form, generate specific task prompts for each medical staff, and the task prompts include: basic patient information, abnormal signs, and urgency; Through the monitoring score of the patient Determine the alert level and alert the corresponding medical staff according to the alert level.

[0016] Furthermore, through the patient's monitoring score Determine the alert level and alert the corresponding medical staff according to the alert level, including: If ≥ 4, the alert level is critically ill, and high-intensity vibration reminder is carried out through the wearable device. If the medical staff does not give feedback, red light fast flashing reminder and high-volume multi-frequency beeping are carried out, and task prompts and monitoring confirmation are displayed on the screen; If 2.5 ≤ < 4, the alert level is high risk, and medium-intensity vibration reminder is carried out through the wearable device. If the medical staff does not give feedback, orange light fast flashing reminder and medium-volume intermittent beeping are carried out, and task prompts and monitoring confirmation are displayed on the screen; If 1.5 ≤ < 2.5, the alert level is medium risk, and low-frequency short vibration reminder is carried out through the wearable device. If the medical staff does not give feedback, yellow light slow flashing and low-volume single prompt are carried out, and task prompts and monitoring confirmation are displayed on the screen; If < 1.5, the alert level is general, and vibration reminder is carried out through the wearable device, and task prompts and monitoring confirmation are displayed on the screen.

[0017] Advantages of the present invention: First, by introducing a multi-parameter dynamic scoring formula for vital signs, combining past medical history correction and weight adaptive mechanism, problems such as difficult objective grading of patient risks, insensitive response to the trend of disease deterioration, and easy omission of key patients are solved, realizing quantitative and refined risk management of patient criticality, greatly improving the identification efficiency of high-risk patients, and enhancing the scientific and intelligent level of monitoring.

[0018] Second, by adopting a semantic adaptation deep model, comprehensive scoring of multi-parameter load and real-time physical location information, problems such as only looking at specialty in medical staff allocation, ignoring load and fatigue, slow response, and unfair allocation are solved, realizing the priority of the most suitable medical staff, dynamic resource balance, the fastest response with the least fatigue and the most suitable professional personnel, and improving the team work efficiency and safety.

[0019] Third, by establishing a multi-parameter linkage matching model of criticality × medical staff status × distance attenuation and introducing an optimization algorithm, problems such as resource waste, allocation conflicts, and inability to give priority to the treatment of critically ill patients caused by single-factor decision-making in traditional allocation are solved, realizing the global optimal medical staff resource scheduling, maximizing the rescue efficiency, ensuring the priority of key patients, and making the whole process traceable.

[0020] Fourthly, through the multi-modal hierarchical warning (vibration, light, sound, screen), step-by-step upgrade and task confirmation mechanisms, problems such as easy omission of alarms, medical staff alarm fatigue, lack of closed-loop traceability of tasks, and low warning transmission efficiency in special scenarios are solved, realizing the reduction of false alarm / missing alarm probability, ensuring the closed-loop disposal of tasks, direct warning to medical staff, and convenient operation, and significantly improving patient safety and information traceability. Brief Description of the Drawings

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0022] Figure 1 It is the flowchart of the method in this embodiment. Detailed Embodiments

[0023] For the convenience of those skilled in the art to understand, the embodiments will be further described in detail below in combination with the accompanying drawings of the structure of the present invention. It should be understood that in the steps mentioned in this embodiment, unless otherwise specified, their sequence can be adjusted according to actual needs, and even can be executed simultaneously or partially simultaneously.

[0024] As Figure 1 shown, the embodiment of the present invention provides a medical monitoring method based on multi-modal prompt interaction, including: S1 Obtain the vital sign data of the patient, score and sort according to the patient information and vital sign data to obtain the patient monitoring score, and form a patient monitoring sequence; S101 Obtain the real-time heart rate data, blood pressure data, blood oxygen saturation data, respiratory rate data, and body temperature of the patient; S102 Through the real-time heart rate data of the patient , blood pressure data , blood oxygen saturation data , respiratory rate data , body temperature calculate the physical sign score of each index of the patient ; The physical sign score formula for the heart rate data in S1021 is: ; The physical sign score formula for the blood pressure data in S1022 is: ; The physical sign score formula for the blood oxygen saturation data in S1023 is: ; The physical sign score formula for the respiratory rate data in S1024 is: ; The physical sign scoring formula for S1025 body temperature data is as follows: ; In this step, a physical sign scoring function for the corresponding index is constructed through the normal range. The median value of the normal range of heart rate, which is 80 in the range of 60 - 100 bpm, is used as the reference point molecule to reflect the actual deviation amplitude. The radius of the normal range (half of the width of 100 - 60 = 40, which is 20) indicates that for every 20 bpm deviation from the median value, 1 point is obtained. 90 mmHg is the critical value for low blood pressure (shock risk threshold), and 50 can cover the common pathological range (for example, when hypertensive crisis ≥ 180 mmHg, 1.8 points are obtained). An oxygen saturation of ≤ 94% triggers a low oxygen alarm. The normal range of respiratory rate is 12 - 20 breaths per minute, and the middle value of 16 is taken as the reference point. The standard human body temperature is 36 - 37.5 °C, and the body temperature is evaluated as too high or too low through 36.5.

[0025] S103 calculates the patient's monitoring score through the physical sign scores of each index of the patient , and the specific calculation formula is as follows: , where is the weight of the corresponding index.

[0026] Furthermore, if the measurement of each index satisfies ∣Δv / Δt∣≥10% for three consecutive times, where Δv is the change amount of the vital sign value and Δt is the time interval, then the weight is adjusted through the weight adjustment formula. The weight adjustment formula is as follows: , where is the width of the normal range of each index.

[0027] In this step, if the heart rate changes from 80 → 100 within 5 minutes (Δv = 20Δv = 20), then the adjustment factor is 1 + 20 / 40 = 1.5, and the weight changes from 0.25 → 0.375. Three consecutive measurements can avoid interference from accidental noise (such as false low blood pressure caused by probe displacement). ∣Δv / Δt∣≥10% ensures that only significant trend changes are responded to. In addition, the designed weight increase limit (such as ≤ 0.5) prevents a single index from monopolizing the score, and the weight decay lower limit (such as ≥ 0.05) avoids index failure.

[0028] Blood oxygen is directly related to tissue oxygenation. Anoxia for 5 minutes can cause irreversible damage, so it has the highest weight; heart rate is related to the risk of cardiac arrest, and the golden rescue time is only 4 - 6 minutes, so it has a secondary weight; respiratory rate drops earlier than blood oxygen, but is easily affected by subjective factors (such as pain), body temperature changes slowly, mostly reflects chronic pathological processes, and lags behind the core pathological process, so it has the lowest weight; the weight corresponding to blood pressure can be dynamically adjusted. For example, if it drops by 20% within 10 minutes, the weight is increased to 0.30. As shown in Table 1, the scoring of each indicator can be designed with the corresponding basic weights as follows: Table 1 Normal ranges and basic weights of each indicator

[0029] S104 Obtain the patient's past medical history data, and mark the vital signs data for intensive care through the patient's past medical history data. The specific calculation formula is as follows: , Among them, is the medical history correction term, is the global weight coefficient affected by the medical history.

[0030] In this step, is the medical history correction term, reflecting the additional risk bonus of a specific medical history to the monitoring requirements, is the global weight coefficient affected by the medical history, with a default value of 1.0 and rising to 1.5 when there is a medical history.

[0031] The calculation formula of the medical history correction term is as follows: , Among them, is the enhancement coefficient of medical history type k for the related sign , is the sign score associated with medical history k, is the baseline risk bonus of medical history k, and n is the total number of medical history types the patient has.

[0032] In this step, the enhancement coefficient of medical history type k for the related sign is based on clinical research evidence, is calculated based on the corresponding sign score formula in the previous step. The baseline risk bonus of medical history k can reflect the default risk of chronic diseases. Through the risk weighting formula of the medical history, the risks of chronic underlying diseases and important past medical histories are truly reflected in the score.

[0033] S2 Obtain the professional ability data, current workload, and location data of medical staff, sort and score according to the professional ability data, current workload, and location data of medical staff to obtain the medical staff status score, and form the medical staff allocation sequence; S201 calculates the adaptation degree A between the professional ability data of medical staff and the patient medical record data through a semantic similarity model; In this step, traditional manual matching and hard rules are inefficient and difficult to reflect the extension of knowledge. It is difficult to judge the compatibility of different professional descriptions (such as "myocardial infarction" and "coronary artery syndrome") only by keywords. When retrieving the professional experience, specialties and other professional ability data of medical staff, for each experience data of medical staff and the medical record type of patients, a semantic similarity model such as a pre-trained BERT model can be used to obtain the adaptation degree A between the professional ability data of medical staff and the patient medical record data, and the value range is generally [0,1]. BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained deep language model based on the Transformer architecture. It can understand and represent complex semantic relationships in natural language and support multi-granularity expressions such as "words, sentences, phrases, and documents". Medical specialties are usually short texts (such as "Department of Cardiology", "Emergency and Critical Care", "Respiratory Diseases"). Case types are also short texts (such as "Myocardial Infarction", "Pulmonary Infection", "Arrhythmia"). BERT can "understand" synonyms, near-synonyms, professional terms, abbreviations, and the meaning conversion between long and short sentences. For example, a high similarity can be obtained between "Myocardial Infarction" and "Acute Coronary Syndrome", that is, even if the words used are different, it can automatically, dynamically and accurately evaluate the professional adaptation degree of each medical staff to a specific case.

[0034] S202 calculates the current workload U of the medical staff according to the current task quantity and working hours of the medical staff; S2021 obtains the current task quantity current_tasks of the medical staff and the maximum task quantity max_capacity that they can carry; S2022 obtains the continuous working time t, and calculates the current workload U through the current task quantity current_tasks of the medical staff, the maximum task quantity max_capacity that they can carry and the continuous working time t. The calculation formula is as follows: ; In this step, the traditional allocation mechanism only considers whether an order can be accepted, without considering the problems of fatigue caused by the cumulative workload of medical staff and the sharp increase in risks after continuous high-intensity work, which is likely to lead to overwork and potential medical safety hazards. When U gets closer to 1, it means the workload is very heavy. It reflects the fatigue spillover caused by long-term work and objectively quantifies the current "busyness / fatigue degree" of medical staff.

[0035] S203 through the moving distance d of the medical staff and the required equipment preparation time Calculate the path cost C from the current position to the patient's position, and the calculation formula is as follows: , where 、 are the weights corresponding to the moving distance d and the required equipment preparation time respectively; In this step, traditionally, the physical arrival efficiency of different medical staff is not dynamically weighted, and the time-consuming of equipment preparation before special operations is rarely considered. Therefore, it may allocate "medical staff with high level but the farthest distance / preparation", resulting in response delays. The moving distance d of medical staff represents the distance between the current position of the medical staff and the position of the patient to be assigned, usually in meters. The farther the distance, the slower the speed of the medical staff to reach the target patient, and the worse the response timeliness. The required equipment preparation time represents the time required to prepare additional medical equipment (such as emergency carts, monitors, infusion pumps, etc.) for completing the current assigned task (such as first aid, specific intervention). If already in the "fully equipped state", this item can be zero. If equipment needs to be carried or configuration is time-consuming, the score increases. The smaller C is, the faster the response speed and the easier it is for the patient to be treated promptly. Improve the first aid / resuscitation response efficiency and adapt to complex sites and clinical scenarios (such as large ICUs, pre-hospital first aid).

[0036] S204 Calculate the medical staff status score according to the data adaptability A, the previous workload U, and the path cost C , and the calculation formula is as follows: ; S205 Form a medical staff allocation sequence from high to low according to the medical staff status score.

[0037] In this step, the larger A is, the larger it means the priority of those with professional fitness and ability; the larger U is, the smaller 1 - U is, a significant decrease indicates fatigue, overload, that is, the priority decreases; the larger C is, the smaller it means the higher the response cost, so the allocation priority is low. The medical staff status score is evaluated by comprehensively considering the adaptability, load level, and the fastest arrival time.

[0038] S3 Select the optimal medical staff from the medical staff allocation queue according to the priority order of the patients in the patient monitoring queue; S301 Select several patients with the highest priority from the patient monitoring sequence to form a sequence of patients to be assigned; S302 For each patient in the sequence of patients to be assigned , traverse all medical staff in the medical staff allocation sequence , calculate the score for each patient-staff combination, and correct the score for each patient-staff combination by distance. , the calculation formula is as follows: , where, is the guardianship score of patient , is the status score of the medical staff , is the physical distance between patient and the medical staff , is the distance attenuation coefficient, is the distance correction factor; In this step, when resources are limited, how to dynamically narrow the scheduling scope that requires the most attention, avoid interference from a large amount of irrelevant data in the whole hospital for peak / emergency task allocation, and ensure the efficiency of batch scheduling and centralized allocation. Traditional medical staff allocation generally only looks at single-factor allocation (such as purely looking at critical conditions, purely looking at distance, purely looking at the status of medical staff, etc.). Through steps S1 and S2, the guardianship score of patient and the status score of the medical staff can be calculated, quantifying multiple parameters. The distance correction factor greatly shortens the response time, ensuring the rescue efficiency and patient safety.

[0039] S303 forms a combined score matrix for each patient-staff combination score, selects the patient-staff combination score that maximizes the global total score, and obtains the allocation table.

[0040] In this step, traditional medical staff allocation only looks at the maximum score or the nearest medical staff, which is likely to lead to individual resource conflicts or omissions. The Hungarian optimization algorithm is used to support one-to-one optimal allocation, maximizing the overall score and achieving the optimal utilization rate of team-level nursing resources.

[0041] S4 generates a multi-modal prompt strategy according to the patient guardianship score, and performs multi-modal prompts on the selected optimal medical staff according to the multi-modal prompt strategy.

[0042] S401 generates specific task prompts for each medical staff according to the allocation table, and the task prompts include: patient basic information, abnormal signs, and urgency; S402 determines the prompt level through the patient's guardianship score , and prompts the corresponding medical staff according to the prompt level; S4021 If ≥4, the prompt level is extremely critical, and high-intensity vibration reminder is performed through the wearable device. If the medical staff does not give feedback, red light fast flashing reminder and high-volume multi-frequency beeping are performed, and the task prompt and guardianship confirmation are displayed on the screen; S4022 if 2.5≤ If the risk is less than 4, the alert level is high-risk, and the wearable device will vibrate with medium intensity to remind you. If the medical staff does not respond, the device will flash an orange light and buzz intermittently at a medium volume, and display the task prompt and monitoring confirmation on the screen. S4023 if 1.5≤ If the value is less than 2.5, the alert level is medium-risk, and the wearable device will give a short low-frequency vibration reminder. If the medical staff does not respond, the yellow light will flash slowly and the low volume will be used for a single reminder. The screen will also display the task reminder and monitoring confirmation. S4024 If the alert level is 1.5, it is a general one, with vibration reminders through the wearable device, and task prompts and monitoring confirmations displayed on the screen.

[0043] In this step, different patient risk stratifications (extremely critical, high risk, medium risk, and general) correspond to different levels of prompts, which can focus limited attention and resources on the people and things that need them most, avoid high-intensity interference from all warnings, prevent warning numbness caused by "alarm fatigue", and ensure that real critical situations will not be ignored. Requiring task prompts to be displayed on the screen and must be confirmed helps record the operation trajectory, ensure the closed loop of alarms, and improve medical safety and quality management.

[0044] Combining the four prompting methods of vibration, light, beep, and screen, we can combat the uncertainty of the on-site environment (such as noise in the operating room, insufficient lighting during night shifts, and the screen is not easy to notice, etc.) to the greatest extent possible. Through multi-sensory stimulation, we can increase the probability of medical staff perceiving alarms in complex scenarios and reduce the missed alarm rate. We first use methods with less impact (such as vibration), and only upgrade to stronger methods when medical staff fail to confirm in time. This "weak first, then strong" strategy not only protects the medical staff's work concentration (without frequent and high-intensity interference), but also ensures that key situations will not be missed.

[0045] Wearable devices can be worn on the wrists of medical staff. When medical staff are busy, walking, or operating, vibration and light signals can be sensed immediately. The skin on the wrist is sensitive, and vibration is highly perceived and visible. Light prompts can be seen by lowering the head or raising the hand. Wearing it on the wrist can facilitate task confirmation, message browsing, and button operation.

[0046] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0047] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0048] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0049] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0050] It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim listing several devices, several of these devices may be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words may be interpreted as names.

[0051] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to cover the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0052] It is obvious that those skilled in the art can make various changes and deformations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and deformations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and deformations.

[0053] In the present invention, unless otherwise clearly defined and limited, the terms "mounted", "connected", "coupled", "fixed", etc. shall be construed in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral body; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0054] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

Claims

1. A medical monitoring method based on multi-modal prompt interaction, characterized in that Including: Obtain the vital sign data of the patient, score and rank according to the patient information and vital sign data to obtain the patient monitoring score, and form a patient monitoring sequence; Obtain the professional ability data, current workload, and location data of the medical staff, sort and score according to the professional ability data, current workload, and location data of the medical staff to obtain the medical staff status score, and form a medical staff allocation sequence; Select the optimal medical staff from the medical staff allocation queue according to the priority order of the patients in the patient monitoring queue; Generate a multimodal prompt strategy according to the patient monitoring score, and perform multimodal prompts on the selected optimal medical staff according to the multimodal prompt strategy.

2. The medical monitoring method based on multi-modal prompt interaction according to claim 1, characterized in that The obtaining the vital sign data of the patient, scoring and ranking according to the patient information and vital sign data to obtain the patient monitoring score, and forming a patient monitoring sequence includes: Obtain the real-time heart rate data, blood pressure data, blood oxygen saturation data, respiratory rate data, and body temperature of the patient; Based on the patient's real-time heart rate data , blood pressure data , blood oxygen saturation data , respiratory rate data , body temperature Calculate the physical sign scores for each indicator of the patient ; Calculate the patient's monitoring score based on the physical signs score of each index of the patient , and the specific calculation formula is as follows: ， Among them, is the weight of the corresponding index; Obtain the patient's past medical history data, and mark the key monitored vital sign data through the patient's past medical history data. The specific calculation formula is as follows: ， Among them, is the medical history correction item, is the global weight coefficient affected by the medical history.

3. The medical monitoring method based on multimodal prompt interaction according to claim 2, wherein, The real-time heart rate data of the patient , blood pressure data , blood oxygen saturation data , respiratory rate data , body temperature Calculate the physical sign score of each index of the patient including: The physical sign scoring formula for heart rate data is ; The physical sign scoring formula for blood pressure data is ; The physical sign scoring formula for blood oxygen saturation data is ; The physical sign scoring formula for respiratory rate data is ; The physical sign scoring formula for body temperature data is .

4. A medical monitoring method based on multimodal prompt interaction according to claim 2, characterized in that, Calculating the patient's monitoring score based on the physical signs scores of each indicator of the patient including: If the three consecutive indicator measurements satisfy |Δv / Δt|≥10%, where Δv is the change in vital sign value and Δt is the time interval, then adjust the weight through the weight adjustment formula. The weight adjustment formula is as follows: ， Among them, is the width of the normal range of each index.

5. The medical monitoring method based on multi-modal prompt interaction according to claim 2, characterized in that The calculation formula of the medical history correction term is as follows: ， Among them, is the enhancement coefficient of the medical history type k for the relevant physical signs , is the physical sign score associated with the medical history k, is the baseline risk addition of the medical history k, and n is the total number of medical history types the patient has.

6. The medical monitoring method based on multimodal prompt interaction according to claim 1, wherein, The obtaining the professional ability data, current workload, and location data of the medical staff, sorting and scoring according to the professional ability data, current workload, and location data of the medical staff to obtain the medical staff status score, and forming a medical staff allocation sequence includes: Calculate the adaptation degree A between the professional ability data of the medical staff and the patient's medical record data through a semantic similarity model; Calculate the current workload U of the medical staff according to the current task quantity and working hours of the medical staff; Through the moving distance d of medical staff and the required equipment preparation time Calculate the path cost C from the current location to the patient location, and the calculation formula is as follows: ， Among them, , are the moving distance d and the required equipment preparation time corresponding weights, respectively. Calculate the status score of medical staff according to the data adaptability A, the previous workload U, and the path cost C , and the calculation formula is as follows: ； The medical staff status scores form a medical staff allocation sequence from high to low.

7. A medical monitoring method based on multimodal prompt interaction according to claim 6, characterized in that The calculating the current workload U of the medical staff according to the current task quantity and working hours of the medical staff includes: Obtain the current task data volume current_tasks of the medical staff and its maximum bearable task quantity max_capacity; Obtain the continuous working time t, and calculate the current workload U through the current task data volume current_tasks of the medical staff, its maximum bearable task quantity max_capacity, and the continuous working time t. The calculation formula is as follows: 。 8. The medical monitoring method based on multi-modal prompt interaction according to claim 1, characterized in that, The selecting the optimal medical staff from the medical staff allocation queue according to the priority order of the patients in the patient monitoring queue includes: Select several patients with the highest priority from the patient monitoring sequence to form a sequence of patients to be allocated; For each patient in the patient sequence to be assigned traverse all medical staff in the medical staff assignment sequence calculate the score for each patient-medical staff combination and correct the score for each patient-medical staff combination by distance The calculation formula is as follows: , Among them, is the guardianship score of the patient , is the status score of the medical staff , is the physical distance between the patient and the medical staff , is the distance attenuation coefficient is the distance correction factor; Form a combined score matrix for each patient-medical staff combination score, and select the patient-medical staff combination score that maximizes the global total score to obtain an allocation table.

9. A medical monitoring method based on multi-modal prompt interaction according to claim 8, characterized in that, The generating a multimodal prompt strategy according to the patient monitoring score, and performing multimodal prompts on the selected optimal medical staff according to the multimodal prompt strategy includes: Generate specific task prompts for each medical staff according to the allocation table, and the task prompts include: basic patient information, abnormal signs, and urgency level; Based on the patient's monitoring score Determine the prompt level and prompt the corresponding medical staff according to the prompt level.

10. A medical monitoring method based on multimodal prompt interaction according to claim 9, characterized in that, The above-mentioned monitoring score of the patient Determining a prompt level and prompting corresponding medical staff according to the prompt level includes: If ≥ 4, the warning level is critically endangered, and a high-intensity vibration reminder is sent through the wearable device. If the medical staff does not give feedback, a fast flashing red light reminder and a high-volume multi-frequency beep are given, and task reminders and guardianship confirmations are displayed on the screen; If 2.5 ≤ <4, it indicates a high-risk level, and medium-intensity vibration reminders are sent through wearable devices. If medical staff do not provide feedback, orange light flashing reminders and medium-volume intermittent beeping are given, and task reminders and guardianship confirmations are displayed on the screen. If 1.5 ≤ <2.5, the prompt level is medium risk, and low-frequency short vibration reminders are given through wearable devices. If medical staff do not give feedback, slow flashing of yellow lights and single low-volume prompts are given, and task prompts and monitoring confirmations are displayed on the screen; If 1.5, the prompt level is normal, with vibration reminders through the wearable device and task prompts and guardianship confirmations displayed on the screen.

Images