A safety monitoring system for use of a medical device of category II

Abstract

The application discloses a kind of second class medical instrument use safety monitoring system, it is related to medical instrument safety technical field, including data acquisition unit, safety evaluation unit and early warning decision unit, data acquisition unit real-time acquisition equipment operating parameter and operating behavior data, wherein operation compliance degree is compared by actual operation step and standard template and is calculated according to integrity, sequentiality, time sequence weighting, safety evaluation unit generates risk index by multidimensional model fusion data, the weight of operation compliance degree in model is increased from 50% to 70% when being lower than threshold value, early warning decision unit is early warning when index exceeds dynamic threshold value, and the threshold value is constituted by benchmark threshold value and the product of three factors of user proficiency, equipment health degree and environmental risk, is positively correlated with historical compliance and maintenance state respectively, and is negatively correlated with environmental risk.The application realizes multidimensional active monitoring, and reduces accident probability.

Description

Technical Field

[0001] This invention relates to the field of medical device safety technology, specifically to a Class II medical device safety monitoring system. Background Technology

[0002] Class II medical devices refer to medical devices whose safety and effectiveness should be controlled. Typical products include blood pressure monitors, electrocardiographs, infusion pumps, ultrasound diagnostic instruments, and medical centrifuges. Their safety is affected by three main factors: the device's own operating status, the user's compliance with operating procedures, and the environmental conditions of use.

[0003] The existing safety monitoring technology for Class II medical devices has the following shortcomings:

[0004] 1. Patent document CN120998529A discloses a method and system for information monitoring of pharmaceutical and medical device data. The document states that it discloses a method and system for information monitoring of pharmaceutical and medical device data, specifically relating to the field of information monitoring of pharmaceutical and medical device data. This includes acquiring the execution request of a pharmaceutical and medical device call task; extracting the preset operation node sequence, execution order, time constraint parameters, and behavioral chain structure corresponding to the pharmaceutical and medical device call task based on the execution request; constructing a hidden trajectory based on the corresponding flow structure information; setting non-interactive markers for the nodes of the hidden trajectory to prevent triggering any device commands and displaying no user interface; and identifying abnormal adjustments, skipped executions, and human intervention during task execution by comparing the execution path structure with the constructed standard hidden trajectory and the real task trajectory, and scoring the consistency of behavior. However, the document focuses on trajectory comparison and anomaly identification of pharmaceutical and medical device call tasks, without real-time fusion analysis of device operating status parameters and operational behavior data, and without constructing a multi-dimensional safety assessment model for Class II medical devices. It cannot dynamically adjust safety thresholds and provide graded early warning responses during operation, thus exhibiting technical deficiencies in real-time monitoring and accuracy of risk response. Summary of the Invention

[0005] The purpose of this invention is to provide a safety monitoring system for the use of Class II medical devices to solve the technical problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a Class II medical device safety monitoring system, wherein the Class II medical device is a medical device used for diagnosis, surgery, or identification, comprising:

[0007] The data acquisition unit collects the operating status parameters of the device and the current user's operational behavior data in real time;

[0008] The safety assessment unit, based on a multi-dimensional safety assessment model, integrates and analyzes the operating status parameters and operational behavior data to generate a safety risk index.

[0009] The early warning decision unit executes an early warning strategy when the safety risk index exceeds the dynamic safety threshold.

[0010] The operational behavior data includes operational compliance, which is obtained by comparing the collected sequence of operational steps with a pre-stored standard operational process template in real time, and by weighting the results according to three dimensions: completeness of operational steps, correctness of sequence, and rationality of timing.

[0011] In the multidimensional safety assessment model, the baseline weight of operational compliance is 50%. When it is lower than the compliance judgment threshold, the weight is smoothly increased to 70%, and the weights of the other parameters are reduced proportionally.

[0012] The dynamic safety threshold is calculated as: baseline safety threshold × user proficiency factor × device health factor × environmental risk correction factor. The user proficiency factor is positively correlated with the user's historical compliance record, the device health factor is positively correlated with the device's maintenance status, and the environmental risk correction factor is negatively correlated with the degree of environmental risk in the current environmental information.

[0013] Preferably, the data acquisition unit includes:

[0014] The operating parameter acquisition subunit is used to acquire at least one of the electrical parameters, functional output parameters, and operating time of the Class II medical device.

[0015] The operation behavior acquisition subunit is used to acquire at least one of the user's operation step sequence, operation duration, and operation posture information for Class II medical devices. The operation behavior acquisition subunit includes a compliance comparison module, which stores a standard operation process template and is used to perform real-time comparison to obtain the operation compliance.

[0016] Preferably, it also includes an environmental sensing unit for collecting the current usage environment information, which includes at least one of ambient temperature, humidity, electromagnetic interference intensity, and lighting conditions. The safety assessment unit is used to incorporate the current usage environment information into a multi-dimensional safety assessment model.

[0017] Preferably, the multidimensional security assessment model is an assessment model based on weighted scoring, fuzzy logic, or machine learning. When a weighted scoring model is used, the operating status parameters, operational compliance, and current usage environment information are respectively assigned benchmark weights. The benchmark weights are dynamically adjusted according to the current risk scenario, and the current risk scenario is determined based on at least one of operational compliance and environmental risk level.

[0018] Preferably, the dynamic security threshold is determined as follows:

[0019] Obtain the user's historical compliance record, and calculate the user's proficiency factor based on the historical compliance record. The better the historical compliance record, the higher the user's proficiency factor.

[0020] The maintenance status of Class II medical devices is obtained, including at least one of the following: cumulative number of uses, time interval since the last maintenance, and remaining life of critical components. Based on this, a device health factor is generated. The better the maintenance status, the larger the device health factor.

[0021] Obtain current environmental information and generate an environmental risk correction factor. The higher the environmental risk, the smaller the environmental risk correction factor.

[0022] The dynamic safety threshold is obtained by multiplying the baseline safety threshold by the user proficiency factor, the device health factor, and the environmental risk correction factor.

[0023] Preferably, the early warning strategy includes a tiered early warning mechanism, and the early warning decision unit is specifically used for:

[0024] When the safety risk index reaches the first warning range, a visual reminder will be issued through the local display interface or indicator light of the Class II medical device.

[0025] When the safety risk index reaches the second warning range, an audible alarm will be activated and a warning notification will be sent to a preset mobile terminal.

[0026] When the safety risk index reaches the third warning interval, a forced interruption command is generated to control the Class II medical device to enter a safety lock state or stop performing its current function.

[0027] The risk levels corresponding to the first, second, and third warning intervals increase sequentially.

[0028] Preferably, it also includes a user authentication unit, which is connected to the data acquisition unit to verify whether the current user is an authorized user, and sends an unauthorized use event to the early warning decision unit when the verification fails, so as to trigger the corresponding early warning strategy.

[0029] Preferably, the dynamic security threshold is adjusted according to the preset permission level of the authorized user, and different permission levels correspond to different operation allowance ranges and different security threshold adjustment coefficients. The higher the permission level, the larger the security threshold adjustment coefficient.

[0030] Preferably, it also includes a security incident tracing unit and a remote monitoring interface;

[0031] The security event tracing unit is used to continuously store operating status parameters and operation behavior data in a circular cache manner, and when the early warning decision unit determines that a security event has occurred, it extracts all cached data within a first preset time period before the security event occurs and a second preset time period after it, and stores them together as an event tracing data packet.

[0032] The remote monitoring interface is used to establish a communication connection with the remote security monitoring platform and upload at least one of the following to the remote security monitoring platform: security risk index, security events, and event traceability data packets.

[0033] Preferably, the security event tracing unit is also used to store snapshots of dynamic security threshold calculations and weight allocation of multidimensional security assessment models corresponding to security events.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1. This invention is equipped with a data acquisition unit, which includes an operating parameter acquisition subunit and an operation behavior acquisition subunit. The operation behavior acquisition subunit is equipped with a compliance comparison module, which can compare the user's actual operation step sequence with a pre-stored standard operation procedure template in real time. The operation compliance is calculated by weighting three dimensions: completeness of steps, correctness of sequence, and rationality of timing. This integrates the device operation status and operation behavior into a unified monitoring system, realizing multi-dimensional and proactive safety risk perception. It can pre-capture potential risks caused by improper operation and significantly reduce the probability of medical accidents.

[0036] 2. This invention is equipped with a safety assessment unit and an early warning decision unit. The safety assessment unit runs a multi-dimensional safety assessment model to generate a safety risk index by fusing and analyzing operating parameters and operational behavior data. The weight of operational compliance in the model can be dynamically adjusted according to its value to focus on the current main sources of risk. The early warning decision unit uses a three-factor product dynamic safety threshold for judgment, so that the threshold can adaptively float according to the state of people, machines and environment. Comparative experiments show that compared with the fixed threshold scheme, the false alarm rate is reduced by about 45% and the false alarm rate is reduced by about 30%, which effectively balances alarm sensitivity and specificity, and avoids medical staff from "alarm fatigue" due to frequent false alarms or missing the opportunity to deal with the situation due to false alarms.

[0037] 3. This invention is equipped with an early warning decision unit, which has a built-in hierarchical early warning mechanism. When the safety risk index exceeds the dynamic safety threshold, it triggers a three-level progressive response based on the degree of risk: visual reminder, sound alarm and remote notification, and forced interruption of operation. This ensures that minor abnormalities are alerted without interrupting the workflow, while serious dangers are blocked instantly, thus protecting patient safety and avoiding excessive interference with normal diagnosis and treatment activities.

[0038] 4. This invention, by installing a user authentication unit, a security event tracing unit, and a remote monitoring interface, enables the user authentication unit to verify the operator's identity and permissions, preventing unauthorized use from the source. The security event tracing unit continuously stores operating parameters and operational behavior data in a circular cache manner. When a security event occurs, it automatically extracts all data from the time periods before and after the event and stores them together as an event tracing data package. At the same time, it saves snapshots of dynamic security threshold calculations and model weight allocation. The remote monitoring interface uploads the security risk index, security events, and tracing data packages to a remote platform, forming a complete "black box" of equipment use, providing reliable data support for post-event liability determination, root cause analysis, and operational training improvements. Attached Figure Description

[0039] Figure 1 This is a structural block diagram of the Class II medical device safety monitoring system of the present invention;

[0040] Figure 2 This is a schematic diagram illustrating the specific composition of the data acquisition unit of the present invention;

[0041] Figure 3 This is a flowchart illustrating the calculation process for dynamically adjusting the safety threshold according to the present invention.

[0042] Figure 4 This is a flowchart illustrating the graded early warning mechanism of the present invention;

[0043] Figure 5 This is a schematic diagram of the dynamic weight adjustment rule of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] System core architecture: The system includes a data acquisition unit, a security assessment unit, an early warning decision unit, as well as an environmental perception unit, a user authentication unit, a security event tracing unit, and a remote monitoring interface that work together. The units communicate with each other through an internal bus or network. The system can be deployed on the embedded main control board of the Class II medical device or in an edge computing gateway that communicates with the device.

[0046] Data acquisition unit: The data acquisition unit includes a sub-unit for acquiring operating parameters and a sub-unit for acquiring operational behavior;

[0047] The operating parameter acquisition subunit acquires real-time operating data of the device through built-in or external sensor modules. The parameters acquired vary for different types of devices. For example, the infusion pump acquires motor speed, tubing pressure, and set flow rate versus actual flow rate; the electrocardiograph acquires lead impedance, common-mode rejection ratio, and signal drift; and the centrifuge acquires rotor speed, vibration amplitude, and motor current. The operating parameters are stored in a time-series format as structured data, and the sampling frequency is set according to the rate of parameter change.

[0048] The operation behavior acquisition subunit captures the user's operation behavior through a combination of multiple sensors. Cameras are installed around the instrument's operation panel to capture the operator's hand movements and key sequence. For large instruments that require specific posture operation, an inertial measurement unit can also be configured to collect operation posture information.

[0049] The compliance comparison module is the core component of the operation behavior collection subunit. This module maintains a standard operation procedure template database. For each model of Class II medical device, there is a corresponding standard operation procedure template. The template is defined using a finite state machine model, which decomposes the correct operation process of the device into a series of ordered state nodes and legal transition conditions between nodes. For example, the standard template for an infusion pump includes state nodes such as "power-on self-test → identity verification → tubing installation confirmation → rate setting → secondary confirmation → start infusion → operation monitoring → infusion completion".

[0050] The compliance comparison module receives the sequence of operation steps captured by the operation behavior collection subunit in real time. It uses a dynamic time warping algorithm to match the actual operation sequence with the status nodes in the standard template. The operation compliance score is calculated from three dimensions (0 to 100 points): step completeness (weight 40%), which is the proportion of completed steps to the number of steps that should be completed; sequence correctness (weight 35%), which is the degree of deviation between the execution order of the steps and the template; and timing rationality (weight 25%), which is whether the operation time of each step is within the normal range. The operation compliance score is input to the security assessment unit at a fixed frequency.

[0051] Environmental sensing unit: The environmental sensing unit consists of a set of environmental sensors, which are installed on or near the Class II medical device body. The sensor set includes a temperature sensor, a humidity sensor, an electromagnetic radiation intensity probe, and a light intensity sensor. Each sensor sends environmental data to the safety assessment unit for comprehensive assessment of the environmental risk level.

[0052] The comprehensive assessment method for environmental risk level is as follows: First, normal range, warning range and danger range are set for each environmental parameter. According to the range in which each parameter is currently located, a corresponding individual risk score is assigned. Then, the individual risk scores are summed according to preset weights to obtain the comprehensive environmental risk score. Finally, the environmental risk level (normal, slight disturbance, moderate disturbance, severe disturbance) is mapped according to the comprehensive environmental risk score.

[0053] Multidimensional security assessment model and dynamic weights: The security assessment unit is the core processing module for multi-source data fusion. It runs a multidimensional security assessment model internally and supports three optional implementation methods: weighted scoring, fuzzy logic, and machine learning.

[0054] Taking the default weighted scoring model as an example, the input parameters are divided into three categories: operational status parameters, operational compliance, and environmental information. Each input parameter is normalized and mapped to a risk contribution score of 0 to 100. The safety risk index is obtained by weighted summation according to the weights. The default configuration of the baseline weights is as follows: 30% for operational status parameters, 50% for operational compliance, and 20% for environmental information.

[0055] The dynamic weight adjustment mechanism is the core feature of the model: the compliance threshold for operational compliance is preset to 80 points. When the operational compliance score is below 80 points, the current risk scenario is determined to be operational risk-driven. The operational compliance weight smoothly increases from 50% to 70% within 5 seconds, the weight of operational status parameters smoothly decreases from 30% to 20%, and the weight of environmental information smoothly decreases from 20% to 10%. When the environmental risk level reaches moderate interference or above, the weight of environmental information increases from 20% to 40%, and the remaining weights are reduced proportionally. When the risk factors subside (operational compliance recovers to above 80 points and environmental risk decreases to below mild interference for 10 seconds), the weights are restored to the default configuration. This mechanism ensures that the model always focuses on the most important source of risk at present.

[0056] If a machine learning model is selected, a model based on a deep neural network or gradient boosting tree will be used. The model will be trained offline using historical data from this instrument model (including labeled samples from normal operation and various abnormal scenarios, with a sample size of no less than 10,000). After deployment, the model parameters will be incrementally updated via online learning.

[0057] Generation of dynamic safety thresholds: The early warning decision unit follows the... Figure 3 The process shown calculates the dynamic security threshold in real time:

[0058] The first step is to calculate the user proficiency factor. This involves retrieving the average compliance score of the current user's past N operations (N is preset to 30) from the historical compliance record database and mapping it using a lookup table. An average compliance score of ≥95% is mapped to 1.15, 85% to 95% to 1.05, 75% to 85% to 1.00, and <75% to 0.90. The better the historical compliance record, the larger the factor value.

[0059] The second step is to calculate the equipment health factor. The cumulative number of uses, the time interval since the last preventive maintenance, and the remaining life of key components are read from the equipment maintenance management database. The comprehensive evaluation is divided into four levels: "excellent", "good", "medium", and "poor", which are mapped to factors of 1.20, 1.00, 0.80, and 0.60, respectively. The better the maintenance status, the larger the factor value.

[0060] The third step is to calculate the environmental risk correction factor. Based on the comprehensive environmental risk level output by the environmental sensing unit, the factor is mapped as follows: normal 1.00, slight disturbance 0.80, moderate disturbance 0.60, severe disturbance 0.40. The higher the environmental risk, the smaller the factor value.

[0061] Step 4: Synthesize the dynamic safety threshold: Dynamic safety threshold = Baseline safety threshold × User proficiency factor × Device health factor × Environmental risk correction factor;

[0062] The baseline safety threshold is a configurable constant with a default value of 60. When the risk increases, the relevant factor decreases, the threshold decreases, and the alarm sensitivity increases. When the risk decreases, the factor increases, the threshold increases, and unnecessary interference is reduced.

[0063] User Identity and Access Management: The user authentication unit supports any combination of fingerprint recognition, near-field communication card swiping, and facial recognition. The system pre-stores an authorized user database. Each record contains the user ID, role, and access level. The correspondence between access level and security threshold adjustment coefficient is as follows: Level 1 (Intern) 0.90, Level 2 (Regular Nurse or Technician) 1.00, Level 3 (Senior Physician) 1.10, Level 4 (Equipment Administrator) 1.20. The final application threshold = dynamic security threshold × adjustment coefficient. The higher the access level, the more lenient the threshold, giving experienced users greater operational freedom.

[0064] Tiered early warning mechanism: The early warning decision unit 30 has a built-in tiered early warning mechanism. The risk levels corresponding to the first, second, and third early warning intervals increase sequentially. The specific intervals are defined as follows: First early warning interval (threshold to threshold × 1.15), second early warning interval (threshold × 1.15 to threshold × 1.40), and third early warning interval (threshold × 1.40 to 100). When the safety risk index reaches the first early warning interval, a visual reminder is issued without interrupting operation. When it reaches the second early warning interval, an audible alarm is activated and an early warning notification is pushed to the preset terminal. When it reaches the third early warning interval, a forced interruption command is generated, and the control device enters a safety lock state.

[0065] Safety event tracing unit and remote monitoring interface: The safety event tracing unit continuously stores operating status parameters and operational behavior data in a circular cache manner. The cache capacity covers at least the most recent 30 minutes of data. When the early warning decision unit determines that a safety event has occurred (triggers the third early warning interval), it extracts all data within a preset time period before and after the event and packages it into an event tracing data package. The data package contains structured fields such as event number, device information, user information, safety risk index sequence, dynamic threshold calculation snapshot, model weight allocation snapshot, operating parameter sequence, and operation compliance sequence. It is stored locally and uploaded to a remote platform.

[0066] The remote monitoring interface supports Ethernet, Wi-Fi, and cellular networks. It uses TLS-encrypted HTTPS and MQTT protocols to upload security risk indices, security event records, and event traceability data packets to the remote security monitoring platform, enabling centralized monitoring of the safety status of all medical devices in the hospital.

[0067] Example 1: Please refer to Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 The present invention provides a Class II medical device safety monitoring system, applied in a medical infusion pump:

[0068] The infusion pump is a Class II medical device used to precisely control the infusion rate. In this embodiment, the monitoring system is deployed on the embedded main control board of the infusion pump. The operation behavior acquisition subunit obtains the sequence of operation steps through the front panel camera and the button event stream. The standard operation process template sets the transition condition between "rate value input → rate secondary confirmation" as follows: the operator must press the confirmation key within a specified time after setting the rate.

[0069] After the nurse (Level 2 access) logs in by swiping her card, the system retrieves her average compliance rate of the last 30 operations, which is 94%. According to the aforementioned lookup table mapping rules, this average compliance rate falls within the range of 85% to 95%, corresponding to a user proficiency factor of 1.05. The infusion pump has accumulated 5260 hours of use and was last maintained 38 days ago. Based on the maintenance status assessment rules, it is comprehensively rated as "Good," corresponding to an instrument health factor of 1.00. The environmental sensing unit 40 detects a temperature of 24℃, humidity of 45%RH, and electromagnetic interference intensity of 0.8V / m. The comprehensive environmental risk assessment is rated as normal, corresponding to an environmental risk correction factor of 1.00. The dynamic threshold = 60 × 1.05 × 1.00 × 1.00 = 63.00, which is 63 after multiplying by the access coefficient of 1.00.

[0070] During the operation, the nurse set the rate to 60mL / h but did not press the confirmation button. The compliance comparison module detected the missing "secondary confirmation" step. The operation compliance was calculated by weighting the three dimensions of step completeness, correct sequence, and reasonable timing. The operation compliance dropped to 78%, which is lower than the compliance judgment threshold of 80 points. The safety assessment unit dynamically increased the operation compliance weight from 50% to 70%. The safety risk index was calculated to be 75, which exceeds the threshold of 63 and is greater than the upper limit of the first warning interval of 63×1.15=72.45. The second warning interval was triggered. The infusion pump screen flashed a warning and a notification was pushed to the nurse station. The head nurse came to the scene to correct the error and avoid potential medication rate errors.

[0071] A 30-day comparative test showed that the false alarm rate of this solution was reduced by about 45% compared with the fixed threshold solution, and no false alarms were found.

[0072] Example 2: Please refer to Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 The present invention provides a Class II medical device safety monitoring system, applied in a portable ultrasound diagnostic instrument:

[0073] Portable ultrasound diagnostic instruments are classified as Class II medical devices and are susceptible to electromagnetic interference. In this embodiment, the system obtains operating parameters through the data interface of the ultrasound diagnostic instrument. The electromagnetic radiation intensity probe is installed near the probe bracket. The average compliance rate of the doctor's (Level 3 access) operations in the past 30 times is 97%. According to the lookup table mapping rules, the average compliance rate is ≥95%, corresponding to a user proficiency factor of 1.15. The device has accumulated 1200 hours of use and has been in use for 15 days since the last maintenance. According to the maintenance status assessment rules, the overall assessment is "excellent", corresponding to a device health factor of 1.20. The initial comprehensive environmental risk assessment is normal, corresponding to an environmental risk correction factor of 1.00. The dynamic threshold = 60 × 1.15 × 1.20 × 1.00 = 82.80, which is 91.08 after multiplying by the access coefficient of 1.10.

[0074] During the examination, the defibrillator in the adjacent area was activated, and the environmental electromagnetic interference suddenly increased to moderate interference. According to the environmental risk level mapping rule, moderate interference corresponds to an environmental risk correction factor of 0.60, and the threshold drops to 49.68 × 1.10 = 54.648. The safety risk index rises to 58, exceeding the threshold, triggering the first warning interval visual alert. The doctor pauses the examination and continues after the interference is eliminated to avoid the risk of misdiagnosis caused by image distortion.

[0075] Example 3: Please refer to Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5This invention provides a Class II medical device safety monitoring system, applied in medical centrifuges for event tracing:

[0076] The medical centrifuge is a Class II medical device under the identification category. The operation behavior is captured by a camera during the sample addition process. The standard template requires symmetrical placement and balancing one by one. The average compliance rate of the intern (Level 1 access) in the past 30 operations is 65%. According to the lookup table mapping rules, the average compliance rate is <75%, corresponding to a user proficiency factor of 0.90. The equipment has accumulated 3500 hours of operation, and the last replacement of the drive belt has exceeded the manufacturer's recommended cycle. According to the maintenance status assessment rules, the overall assessment is "medium", corresponding to a device health factor of 0.80. The overall environmental risk assessment is normal, corresponding to an environmental risk correction factor of 1.00. The dynamic threshold = 60 × 0.90 × 0.80 × 1.00 = 43.20, which is 38.88 after multiplying by the access coefficient of 0.90.

[0077] During operation, the intern placed four tubes into one side of the rotor without balancing. The compliance comparison module calculated the operation compliance from three dimensions and it plummeted to 35%. The model dynamically increased the operation compliance weight to 70%, and the safety risk index reached 95, far exceeding the threshold of 38.88. The system triggered the third warning interval and forced interruption, locked the motor and issued an alarm.

[0078] The security incident tracing unit extracts data from 30 seconds before and after an incident, including rotor image sequences, compliance change curves, threshold calculation snapshots, and weight allocation snapshots. This data is packaged into an incident tracing data package and uploaded to a remote platform. Managers can then replay the entire incident through the platform and conduct targeted training accordingly, thus forming a closed-loop management system.

[0079] The system operates by installing a data acquisition unit, which includes an operational parameter acquisition subunit and an operational behavior acquisition subunit. The operational behavior acquisition subunit contains a compliance comparison module that compares the user's actual operational steps with a pre-stored standard operating procedure template in real time. It calculates the operational compliance based on three weighted dimensions: completeness of steps, correctness of sequence, and rationality of timing. This integrates the device's operational status and operational behavior into a unified monitoring system, achieving multi-dimensional and proactive safety risk perception. It can proactively detect potential risks caused by improper operation, significantly reducing the probability of medical accidents. Furthermore, the system includes a safety assessment unit and an early warning decision unit. The safety assessment unit uses a multi-dimensional safety assessment model to fuse and analyze operational parameter and operational behavior data, generating a safety risk index. The operational compliance weight in the model can be dynamically adjusted to focus on the main sources of current risk. The early warning decision unit uses a three-factor product dynamic safety threshold for judgment, allowing the threshold to adaptively fluctuate according to the state of the person, machine, and environment. Comparative experiments show that compared to a fixed threshold scheme, the false alarm rate is reduced by approximately 45%, and the false negative rate is reduced by approximately 30%, effectively balancing alarm sensitivity and specificity, and preventing medical personnel from making mistakes. To prevent "alarm fatigue" caused by frequent false alarms or missed opportunities for intervention due to missed alarms, an early warning decision unit is installed. This unit has a built-in hierarchical early warning mechanism. When the safety risk index exceeds the dynamic safety threshold, it triggers a three-tiered progressive response based on the degree of risk: visual alerts, audible alarms and remote notifications, and forced operation interruption. This ensures that minor abnormalities are alerted without interrupting the workflow, while serious dangers are instantly blocked, protecting patient safety and avoiding excessive interference with normal medical activities. The unit also includes a user authentication unit, a safety event tracing unit, and a remote monitoring interface. The user authentication unit verifies the operator's identity and permissions, preventing unauthorized use from the source. The safety event tracing unit continuously stores operating parameters and operational behavior data in a circular cache. When a safety event occurs, it automatically extracts all data from the time periods before and after the event and stores them together as an event tracing data package. It also saves snapshots of dynamic safety threshold calculations and model weight allocation. The remote monitoring interface uploads the safety risk index, safety events, and tracing data packages to a remote platform, forming a complete "black box" of equipment usage. This provides reliable data support for post-event liability determination, root cause analysis, and operational training improvements.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. A safety monitoring system for the use of Class II medical devices, wherein the Class II medical devices are medical devices used for diagnosis, surgery, or identification, characterized in that, include: The data acquisition unit collects real-time operating status parameters of Class II medical devices and current user operation behavior data; The safety assessment unit generates a safety risk index by integrating and analyzing operational status parameters and operational behavior data based on a multi-dimensional safety assessment model. The early warning decision unit executes an early warning strategy when the safety risk index exceeds the dynamic safety threshold. The operational behavior data includes operational compliance, which is obtained by comparing the collected sequence of operational steps with a pre-stored standard operational process template in real time, and by weighting the results according to three dimensions: completeness of operational steps, correctness of sequence, and rationality of timing. In the multidimensional safety assessment model, the baseline weight of operational compliance is 50%. When the operational compliance is lower than the compliance judgment threshold, this weight is smoothly increased to 70%, and the weights of parameters other than operational compliance are reduced proportionally. The dynamic safety threshold is calculated as: baseline safety threshold × user proficiency factor × device health factor × environmental risk correction factor. The user proficiency factor is positively correlated with the user's historical compliance record, the device health factor is positively correlated with the device's maintenance status, and the environmental risk correction factor is negatively correlated with the degree of environmental risk in the current environmental information.

2. The Class II medical device safety monitoring system according to claim 1, characterized in that, The data acquisition unit includes: The operating parameter acquisition subunit is used to acquire at least one of the electrical parameters, functional output parameters, and operating time of the Class II medical device. The operation behavior acquisition subunit is used to acquire at least one of the user's operation step sequence, operation duration, and operation posture information for Class II medical devices. The operation behavior acquisition subunit includes a compliance comparison module, which stores a standard operation process template and is used to perform real-time comparison to obtain the operation compliance.

3. The Class II medical device safety monitoring system according to claim 1, characterized in that: It also includes an environmental sensing unit for collecting the current usage environment information, which includes at least one of ambient temperature, humidity, electromagnetic interference intensity, and lighting conditions. The safety assessment unit is used to incorporate the current usage environment information into a multi-dimensional safety assessment model.

4. The Class II medical device safety monitoring system according to claim 1, characterized in that: The multidimensional security assessment model is an assessment model based on weighted scoring, fuzzy logic, or machine learning. When a weighted scoring model is used, the operating status parameters, operational compliance, and current usage environment information are assigned benchmark weights, which are dynamically adjusted according to the current risk scenario. The current risk scenario is determined based on at least one of operational compliance and environmental risk level.

5. The Class II medical device safety monitoring system according to claim 1, characterized in that, The dynamic security threshold is determined as follows: Obtain the user's historical compliance record, and calculate the user's proficiency factor based on the historical compliance record. The better the historical compliance record, the higher the user's proficiency factor. The maintenance status of Class II medical devices is obtained, including at least one of the following: cumulative number of uses, time interval since the last maintenance, and remaining life of critical components. Based on this, a device health factor is generated. The better the maintenance status, the larger the device health factor. Obtain current environmental information and generate an environmental risk correction factor. The higher the environmental risk, the smaller the environmental risk correction factor. The dynamic safety threshold is obtained by multiplying the baseline safety threshold by the user proficiency factor, the device health factor, and the environmental risk correction factor.

6. The Class II medical device safety monitoring system according to claim 1, characterized in that, The early warning strategy includes a tiered early warning mechanism, and the early warning decision unit is specifically used for: When the safety risk index reaches the first warning range, a visual reminder will be issued through the local display interface or indicator light of the Class II medical device. When the safety risk index reaches the second warning range, an audible alarm will be activated and a warning notification will be sent to a preset mobile terminal. When the safety risk index reaches the third warning interval, a forced interruption command is generated to control the Class II medical device to enter a safety lock state or stop performing its current function. The risk levels corresponding to the first, second, and third warning intervals increase sequentially.

7. The Class II medical device safety monitoring system according to claim 1, characterized in that: It also includes a user authentication unit, which is connected to the data acquisition unit to verify whether the current user is an authorized user, and sends an unauthorized use event to the early warning decision unit when the verification fails, so as to trigger the corresponding early warning policy.

8. The Class II medical device safety monitoring system according to claim 1, characterized in that: The dynamic security threshold is adjusted according to the preset permission level of the authorized user, and different permission levels correspond to different operation allowance ranges and different security threshold adjustment coefficients. The higher the permission level, the larger the security threshold adjustment coefficient.

9. The Class II medical device safety monitoring system according to claim 1, characterized in that, It also includes a security incident tracing unit and a remote monitoring interface; The security event tracing unit is used to continuously store operating status parameters and operation behavior data in a circular cache manner, and when the early warning decision unit determines that a security event has occurred, it extracts all cached data within a first preset time period before the security event occurs and a second preset time period after it, and stores them together as an event tracing data packet. The remote monitoring interface is used to establish a communication connection with the remote security monitoring platform and upload at least one of the following to the remote security monitoring platform: security risk index, security events, and event traceability data packets.

10. The Class II medical device safety monitoring system according to claim 9, characterized in that: The security event tracing unit is also used to store snapshots of dynamic security threshold calculations and weight allocation of multidimensional security assessment models corresponding to security events.

Images