Medical equipment risk management system and method based on data analysis

Abstract

The invention relates to the field of medical equipment risk management, in particular to a medical equipment risk management system and method based on data analysis, and the system comprises an information acquisition and storage module, a data preprocessing module, an internal and external factor association module, an anomaly detection module, a risk assessment module and a decision management module. A total data set is obtained by collecting and storing medical equipment and operation data, and data preprocessing is performed to reduce noise interference; according to the preprocessed medical equipment data information, constructing an anomaly detection neural network model, and outputting a medical equipment anomaly detection result; calculating the correlation degree of the influence of the internal factors and the external factors of the medical equipment on the risk; risk conditions existing during operation of the medical equipment are evaluated in real time, abnormal behaviors of the equipment are found in time, and measures are taken to avoid equipment faults and accidents according to the risk evaluation conditions. According to the invention, effective data analysis and risk assessment can be carried out, and the safety of medical equipment is improved.

Description

Technical Field

[0001] This invention relates to the field of medical device risk management, and more particularly to a medical device risk management system and method based on data analysis. Background Technology

[0002] The safety and reliability of medical devices are crucial to patient health and the operation of healthcare institutions. Medical device risk management refers to monitoring, assessing, and managing device risks to reduce equipment failures and accidents, ensuring the normal operation of equipment and patient safety. Meanwhile, medical devices generate a large amount of data during operation, including operating status, performance parameters, usage, and more. This data can be collected in real time through device interfaces or sensors to monitor the device's operation and performance.

[0003] Traditional medical device risk management methods mainly rely on post-incident reporting and manual review, but they cannot monitor the risk status of equipment in real time, which can easily lead to delayed detection and handling of equipment failures and accidents. With the advancement of technology, modern medical devices generate a huge amount of complex data, containing various types of data. Processing this data and extracting useful information from it is a challenge. The inability to accurately identify the risk status of medical devices poses safety hazards and can cause a certain degree of loss.

[0004] Therefore, it is necessary to design a medical device risk management system and method based on data analysis. Summary of the Invention

[0005] This invention provides a data analysis-based medical device risk management system and method, aiming to improve the safety and reliability of medical devices through effective data analysis and risk assessment, and to identify potential risks and take corresponding measures for management and prevention.

[0006] The specific technical solution of this invention is as follows:

[0007] A data-driven medical device risk management system includes the following components:

[0008] The module includes: information acquisition and storage module, data preprocessing module, internal and external factor correlation module, anomaly detection module, risk assessment module, and decision management module.

[0009] The information acquisition and storage module collects and stores data related to medical equipment through equipment sensors, equipment monitoring systems, and hospital information systems.

[0010] The data preprocessing module is used to preprocess the raw data through cleaning, correction and standardization. It uses smoothing adjustment to reduce the impact of noise interference and ensure the accuracy and consistency of the data.

[0011] The internal and external factors correlation module analyzes the internal and external factors of medical equipment and determines the correlation between these factors and equipment risk.

[0012] The anomaly detection module is used to analyze medical device data and model it to detect anomalies in medical devices in order to identify potential risks.

[0013] The risk assessment module evaluates the risk status of medical equipment and determines its risk level based on abnormal detection results and the correlation between internal and external factors and equipment risk.

[0014] The decision management module is used to provide decision support based on risk assessment results, and helps relevant organizations and personnel take risk management and control measures by generating risk reports, alerts and recommendations.

[0015] A data-driven approach to medical device risk management includes the following steps:

[0016] S1. By collecting and storing the medical equipment itself and its operational data, a total data set is obtained, and data preprocessing is performed to reduce noise interference;

[0017] S2. Based on the preprocessed medical equipment data, an anomaly detection neural network model is constructed. The anomaly detection neural network model can output the anomaly detection results of medical equipment according to the characteristics and historical data of the medical equipment; and calculate the correlation between the internal and external factors of the medical equipment and their impact on risk.

[0018] S3. Conduct real-time assessments of the risks associated with the operation of medical equipment, promptly detect abnormal behavior of the equipment, and take measures based on the risk assessment results to prevent equipment malfunctions and accidents.

[0019] Furthermore, step S2 specifically includes:

[0020]

[0021] Where Mco represents the correlation between internal and external factors of a medical device and its risk; cov(a, b) represents the correlation between the a-th accessory and the v-th type of the medical device, where a∈[1, n], b∈[1, w]; α，1 Indicates the number of times medical equipment needs maintenance; I a，2 Indicates the number of times medical equipment needs to be repaired; β1, β2, and β3 represent the corresponding correlation parameters; O b，1 Indicates operator error; O b，2 This indicates the number of power supply defects.

[0022] Furthermore, step S2 specifically includes:

[0023] A neural network model for detecting anomalies in medical devices is established using deep learning algorithms. The model includes an input layer, a detection layer, a confirmation layer, and an output layer.

[0024] Furthermore, the q attribute information of the medical device data is input into the input layer, represented as: ip∈[1,q], the input layer and the detection layer are fully connected. Anomaly detection is performed on the features of medical device data in the detection layer. The specific process is as follows:

[0025] INU2=ω1INU1+b1

[0026]

[0027] Where INU2 represents the input in the detection layer; TOU2 represents the output of the detection layer; ω1 represents the connection weight between the input layer and the detection layer; b1 represents the bias of the detection layer; λ represents the fusion coefficient; θ represents a constant value to ensure the stability of the calculation process; α1 and α2 represent the corresponding adjustment parameters; INU 1，t This represents the average fluctuation of medical device data obtained at time t; INU 2，t α represents the standardized value of the detection layer input obtained at time t; α represents the activation function; T represents the bias; σ represents the detection parameters; and ca0 represents the preset detection parameter value.

[0028] Furthermore, step S3 specifically includes:

[0029] Based on the output of the medical equipment anomaly detection neural network model and the correlation between the internal and external factors of medical equipment and their impact on the risk of medical equipment, the evaluation coefficient Rs for medical equipment risk management is calculated, defining ip internal factors and ep external factors.

[0030] Furthermore, the risk of medical equipment is managed based on the assessment coefficient. Rs′ is defined as a preset standard assessment parameter, τ represents the system assessment characteristic parameter, and Ran represents the risk level. If Rs′≥Rs, This indicates that the medical equipment is at high risk; if 0 ≤ Rs' < Rs, This indicates that the medical equipment is at medium risk; all other situations indicate that the medical equipment is at low risk.

[0031] Beneficial effects:

[0032] 1. This invention, when preprocessing medical device data, obtains a stable sequence by differencing the unstable time series, transforming the unstable time series into a stable one, making it easier to capture trends and patterns within the sequence. Differential processing also reduces data complexity, providing a data foundation for subsequent modeling. To obtain more accurate data, a Kalman filter is used for state assessment. By considering the system's measurement noise and model uncertainty to suppress their influence, adaptive estimation of measurement and model errors is performed, providing more accurate state assessment results. This invention enables continuous state assessment in medical device monitoring and control, providing real-time, continuous state information. The fusion of multiple data sources improves the accuracy and robustness of state estimation.

[0033] 2. This invention calculates the correlation between internal and external factors and the risk of medical devices, monitors changes in relevant factors, formulates corresponding improvement plans, and takes early measures to mitigate risks and improve the reliability and safety of the equipment. Correlation analysis can provide feedback and guidance for medical device risk management, helping to continuously improve and optimize management measures. By comprehensively considering the interrelationships between multiple factors, the overall risk status of the equipment can be assessed more accurately, providing a more comprehensive basis for risk management. The calculated correlation provides a basis for continuous improvement and tracking of medical device risk management, and lays the groundwork for a comprehensive assessment of medical device risk management.

[0034] 3. This invention establishes an anomaly detection neural network model to discover abnormal behaviors in the operation of medical equipment, automatically monitors and diagnoses the operating status of medical equipment, reduces the burden of manual monitoring, and improves work efficiency. Abnormal states of medical equipment may have a negative impact on patient safety and treatment outcomes. By detecting equipment anomalies in a timely manner, potential risks to patients can be reduced, improving medical quality and safety. Furthermore, it provides data support for subsequent medical equipment risk assessment.

[0035] 4. Based on risk assessment results, the system automatically implements corresponding risk management measures. Through the data analysis process described in this invention, it provides decision support, helping medical personnel select appropriate repair, component replacement, and equipment calibration measures to reduce equipment risk and improve equipment safety and reliability. It can also generate risk reports, alerts, and recommendations to assist relevant institutions and personnel in taking appropriate risk management and control measures. The decision management module also provides monitoring and tracking functions to ensure the effective implementation and evaluation of risk management measures. Attached Figure Description

[0036] Figure 1 This is a flowchart of the data analysis-based medical device risk management method described in this invention;

[0037] Figure 2This is a block diagram of the data analysis-based medical device risk management system described in this invention. Detailed Implementation

[0038] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. It should also be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the scope of the invention.

[0039] See attached document Figure 1 This embodiment provides a data analysis-based method for medical device risk management, including the following steps:

[0040] S1. By collecting and storing the medical equipment itself and its operational data, a total data set is obtained, and data preprocessing is performed to reduce noise interference;

[0041] Various data from medical devices are collected and recorded through interfaces and sensors, equipment monitoring systems, and hospital information systems, providing data support for subsequent analysis. Modern medical devices generally have interfaces and sensors that can collect real-time operational data, including temperature, pressure, current, vibration, usage frequency, and equipment status.

[0042] Simultaneously, a data acquisition and storage mechanism is established to ensure reliable acquisition and persistent storage of device data. The system can use dedicated data acquisition equipment or integrate with interfaces from equipment manufacturers to transmit device data to a central database or cloud platform for storage and management.

[0043] Data related to internal and external factors is collected through interfaces and sensors of medical devices, as well as daily records, to obtain a total data set X, where X = {X...} n X w}; where X n X represents the internal factors of medical devices. w External factors related to medical equipment; 'a' represents the number of components in the medical device, a∈[1,n], X n Any subset of can be represented as The information set representing the a-th component; the internal factors of medical equipment include design defects, substandard production quality, maintenance frequency, repair frequency, and other relevant information data.

[0044] b represents the number of types of medical devices, b∈[1, w], X w Any subset of can be represented as The information set represents the b-th type of medical device; external factors of medical devices include environmental conditions, power supply, number of operator errors, and other relevant information data.

[0045] Based on the total data set obtained from the medical devices, preprocessing is necessary because interfaces or sensors may be subject to noise interference. To ensure data accuracy, parameters related to both internal and external factors of the medical devices need to be assessed to determine whether their time series are stable.

[0046] In a specific embodiment, the internal and external factors of the medical device are represented by c. m This indicates that, in order to ensure the relevant parameters of the medical equipment, namely c m Whether the time series itself is stable needs to be tested for cointegration. If the time series of relevant parameters of one or more medical devices are unstable, they need to be differencing. The specific process is as follows:

[0047]

[0048] in, This represents the difference, used to eliminate instability. Let z1, z2, ..., zn be the recorded values ​​of the relevant parameters of the m-th medical device at time step 0; j0 represents the constant term and the intercept in the autoregression; τ represents the constant value; z1, z2, ..., zn are the values ​​of the parameters of the m-th medical device at time step 0. v The coefficients representing the difference terms are used to describe the short-term dynamic relationships of the time series; v represents the total number of parameters related to medical devices in the time series. The error term represents the random fluctuations during the difference process; s represents the average value of the changing trend of the relevant parameters of the medical equipment in the time series.

[0049] Based on the VECM (Vector Error Correction Model) and a cointegration model, the long-term equilibrium relationship between relevant parameters of medical devices is estimated. The cointegration model can be used to capture the long-term relationship between non-stationary time series, specifically:

[0050]

[0051] in, Par1 represents the change in non-stationary time series variables related to medical equipment parameters; Par2 is a parameter matrix representing the short-term dynamic relationship between multiple non-stationary time series in medical equipment parameters; Par2 is also a parameter matrix representing the correction of the long-term equilibrium relationship by multiple non-stationary time series in medical equipment parameters.

[0052] Then, a residual test was performed, revealing that the residual sequence was stable, i.e., a cointegration relationship existed. Therefore, more accurate data could be obtained, and the Kalman filter could be used to provide a more reliable model for more accurate state assessment in the presence of noise and uncertainty.

[0053] State transition equation: g(o)=F(o-1)*g(o-1)+B(o-1)*u(o-1)+d(0-1)

[0054] Where g(o) represents the state vector at time 0; F(o-1) represents the state transition matrix, which shows how the state vector evolves over time; B(o-1) represents the control input matrix, which shows the influence of external inputs on the state vector; u(o-1) represents the control input vector, which shows the external input at time 0-1; and d(0-1) represents the process noise.

[0055] Observation equation: q(o)=H(o)*g(o)+k(o)

[0056] Where q(o) represents the observation vector at time 0; H(o) represents the observation matrix, which is the relationship between the observation vector and the state vector; and k(o) represents the observation noise.

[0057] The state vector and state covariance matrix are initialized, and then the prediction steps are as follows:

[0058] g hat (o)=F(o-1)*g(o-1)+B(o-1)*u(o-1)

[0059] p hat (o)=F(o-1)*p(o-1)F(o-1) T +Q(o-1)

[0060] Among them, g hat (o) represents the predicted value of the medical device-related parameters at time o; P hat (o) represents the predicted state covariance; p(o-1) represents the estimated state covariance matrix at o-1; T represents the time interval; Q(o-1) represents the process noise covariance matrix.

[0061] The update steps are as follows:

[0062]

[0063] g(o)=g hat (o)+k(o)*(q(o)-H(o)*g hat (o))

[0064] p(o)=(Ik(o)*H(o))*p hat (o)

[0065] Where k(o) represents the gain matrix, the uncertainty trade-off between observed data and predicted values; R(o) represents the observation noise covariance matrix; g(o) represents the more accurate state estimate; p(o) represents the filtered state covariance matrix; I represents the identity matrix; and T represents the time interval.

[0066] This invention obtains a stable sequence by differentiating unstable time series when preprocessing medical device data. The difference operation can eliminate the trend and seasonal components in the time series, making the series more stable and transforming unstable time series into stable ones, making it easier to capture trends and patterns in the series. The difference processing can also reduce the complexity of the data and provide a data foundation for the subsequent modeling process.

[0067] To obtain more accurate data, a Kalman filter is used for state assessment. By considering the measurement noise and model uncertainty of the system to suppress their effects, the measurement and model errors are adaptively estimated and incorporated into the state assessment process, thereby providing more accurate state assessment results. This invention can perform continuous state assessment in medical device monitoring and control, providing real-time and continuous state information. By fusing multiple data sources, the accuracy and robustness of state estimation are improved.

[0068] S2. Based on the preprocessed medical equipment data, construct an anomaly detection neural network model. The anomaly detection neural network model can output the anomaly detection results of medical equipment according to the characteristics and historical data of the medical equipment; and calculate the correlation between the internal and external factors of the medical equipment and their impact on risk.

[0069] S201. Using the preprocessed dataset of internal and external factors of the medical equipment, calculate the correlation between the internal and external factors and their impact on the risk of the medical equipment. The specific process is as follows:

[0070]

[0071] Where Mco represents the correlation between the internal and external factors of a medical device and its risk impact; cov(a, b) represents the correlation between the a-th accessory and the b-th type of the medical device; I a，1 Indicates the number of times medical equipment needs maintenance; I a，2 Indicates the number of times medical equipment needs to be repaired; β1, β2, and β3 represent the corresponding correlation parameters; O b，1 Indicates operator error; O b，2 This indicates the number of power supply defects.

[0072] This invention calculates the correlation between internal and external factors and the risk of medical devices, monitors changes in relevant factors, formulates corresponding improvement plans, and takes early measures to mitigate risks and improve the reliability and safety of equipment. Correlation analysis can provide feedback and guidance for medical device risk management, helping to continuously improve and optimize management measures. By comprehensively considering the interrelationships between multiple factors, the overall risk status of the equipment can be more accurately assessed, providing a more comprehensive basis for risk management. The calculated correlation provides a basis for continuous improvement and tracking of medical device risk management, and lays the groundwork for a comprehensive assessment of medical device risk management.

[0073] S202. By extracting useful information from the operational data of medical equipment, an anomaly detection neural network model is constructed. The model can output anomaly detection results of medical equipment based on the characteristics and historical data of the medical equipment, identify potential risk factors, and predict the possibility of medical equipment failure. The establishment of the model can help the system more accurately assess the risk level of the equipment.

[0074] Based on the acquired total data set of medical devices, a neural network model for anomaly detection of medical devices is established using a deep learning algorithm. The total dataset X of medical devices is represented as... in, This represents the q-th attribute information of the r-th data. The neural network model for anomaly detection in medical devices includes an input layer, a detection layer, a confirmation layer, and an output layer.

[0075] Inputting q attribute information of any medical device data into the input layer, represented as: ip∈[1,q], the input layer and the detection layer are fully connected. Anomaly detection is performed on the features of medical device data in the detection layer. The specific process is as follows:

[0076] INU2=ω1INU1+b1

[0077]

[0078] Where INU2 represents the input in the detection layer; TOU2 represents the output of the detection layer; ω1 represents the connection weight between the input layer and the detection layer; b1 represents the bias of the detection layer; λ represents the fusion coefficient; θ represents a constant value to ensure the stability of the calculation process; α1 and α2 represent the corresponding adjustment parameters; INU 1，t This represents the average fluctuation of medical device data obtained at time t; INU 2，t α represents the standardized value of the detection layer input obtained at time t; α represents the activation function; T represents the bias; σ represents the detection parameters; and ca0 represents the preset detection parameter value.

[0079] In the confirmation layer, abnormal situations are further confirmed to reduce false alarms and risks. The specific process is as follows:

[0080] INU3=ω2TOU2+b2

[0081]

[0082]

[0083] Where INU3 represents the input to the confirmation layer; ω2 represents the connection weight between the detection layer and the confirmation layer; b2 represents the bias of the confirmation layer; TOU3 represents the output of the confirmation layer; P(y l ) represents the probability of a medical device malfunctioning, y l J represents the l-th item of the characteristic medical device data; J represents (y l The transpose of -μ1); j represents the number of features; μ1 represents the mean vector of the features; t 0 t represents the start node of the selected time period; t′ represents the end node of the selected time period; E(t) represents the working energy of the medical device recorded at time t; ε represents the threshold, which is derived based on expert experience and a large number of preliminary experiments.

[0084] If the probability of the medical device reaching the threshold is greater than or equal to the threshold ε, the output result is -1, indicating that the medical device faces a significant risk. The system will immediately issue an alarm, and the medical device administrator will take timely emergency measures. If the probability of the medical device reaching the threshold is less than the threshold ε, the output result is... The confirmation layer continues to output the results.

[0085] Finally, the confirmation layer transmits the abnormal detection results of the medical device to the output layer for output, and further assesses the risk of the medical device based on the output results.

[0086] This invention establishes an anomaly detection neural network model to discover abnormal behaviors in the operation of medical devices, automatically monitors and diagnoses the operating status of medical devices, reduces the burden of manual monitoring, and improves work efficiency. Abnormal states of medical devices may have a negative impact on patient safety and treatment outcomes. By detecting device anomalies in a timely manner, potential risks to patients can be reduced, improving medical quality and safety. Furthermore, it provides data support for subsequent medical device risk assessments.

[0087] S3. Conduct real-time assessments of the risks associated with the operation of medical equipment, promptly detect abnormal behavior of the equipment, and take timely measures based on the risk assessment to prevent equipment failures and accidents.

[0088] Based on the output of the neural network model for anomaly detection in medical equipment and the correlation between the internal and external factors of medical equipment and their impact on risk, the evaluation coefficient Rs for risk management of medical equipment is calculated. There are ip internal factors and ep external factors. The specific process is as follows:

[0089]

[0090] Where, ω ip This represents the weight of each medical device's internal factors influencing its risk; ω ep The weight θ represents the degree to which external factors affect the risk of each medical device. ip The weighted percentage of the risk impact of internal factors of medical equipment; θ ep σ represents the weighting percentage of the risk impact of external factors on medical equipment; σ represents the adjusted parameter value of the evaluation coefficient; Sec represents the risk and safety score of the medical equipment; Sec′ represents the total risk and safety score of the medical equipment.

[0091] The risk of medical equipment is managed based on an assessment coefficient. Rs′ is defined as a preset standard assessment parameter, τ represents the system assessment characteristic parameter, and Ran represents the risk level. If Rs′ ≥ Rs... This indicates that the medical equipment is at high risk; if 0 ≤ Rs' < Rs, This indicates that the medical equipment is at medium risk; all other situations indicate that the medical equipment is at low risk.

[0092] Based on the risk level of the medical equipment obtained from the assessment, corresponding risk management measures are determined. High-risk equipment requires more frequent maintenance and repair, enhanced monitoring and early warning systems, increased spare parts inventory, and improved operator skills. Medium-risk equipment requires regular maintenance and inspection to ensure normal equipment operation and reduce potential risks. Low-risk equipment requires routine maintenance and management measures.

[0093] Based on risk assessment results, this invention automatically implements corresponding risk management measures. Through the data analysis process described in this invention, it provides decision support, helping medical personnel select appropriate repair, component replacement, and equipment calibration measures to reduce equipment risk and improve equipment safety and reliability. It can also generate risk reports, alerts, and recommendations to assist relevant institutions and personnel in taking appropriate risk management and control measures. The decision management module also provides monitoring and tracking functions to ensure the effective implementation and evaluation of risk management measures.

[0094] See attached document Figure 2This embodiment provides a data analysis-based medical device risk management system, including the following:

[0095] The module includes: information acquisition and storage module, data preprocessing module, internal and external factor correlation module, anomaly detection module, risk assessment module, and decision management module.

[0096] The information acquisition and storage module collects and stores data related to medical equipment through equipment sensors, equipment monitoring systems, hospital information systems, etc., including equipment specifications, usage records, maintenance records, fault reports, etc.

[0097] The data preprocessing module is used to preprocess the raw data through cleaning, correction and standardization steps. It uses smoothing adjustment to reduce the impact of noise interference, ensure the accuracy and consistency of the data, and provide a reliable data foundation for subsequent modeling.

[0098] The internal and external factors correlation module analyzes the internal and external factors of medical equipment and determines the correlation between these factors and equipment risk; it also lays the groundwork for subsequent assessment and analysis.

[0099] The anomaly detection module is used to analyze medical device data and model it to detect anomalies in medical devices in order to identify potential risks.

[0100] The risk assessment module evaluates the risk status of medical equipment and determines its risk level based on abnormal detection results and the correlation between internal and external factors and equipment risk.

[0101] The decision management module provides decision support based on risk assessment results. It helps relevant organizations and personnel take appropriate risk management and control measures by generating risk reports, alerts, and recommendations. The decision management module also provides monitoring and tracking functions to ensure the effective implementation and evaluation of risk management measures.

[0102] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0103] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0104] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0105] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0106] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A medical device risk management system based on data analysis, characterized in that, Includes the following parts: The module includes: information acquisition and storage module, data preprocessing module, internal and external factor correlation module, anomaly detection module, risk assessment module, and decision management module. The information acquisition and storage module collects and stores data related to medical equipment through device sensors, device monitoring systems, and hospital information systems. The data preprocessing module is used to preprocess the raw data through cleaning, correction and standardization, and to reduce the impact of noise interference by smoothing adjustment, so as to ensure the accuracy and consistency of the data. The internal and external factors correlation module analyzes the internal and external factors of medical equipment and determines the correlation between these factors and equipment risk. The anomaly detection module is used to analyze medical device data and model it to detect anomalies in medical devices in order to identify potential risks. The risk assessment module assesses the risk status of the medical equipment and determines its risk level based on the abnormal detection results of the medical equipment and the correlation between the internal and external factors of the medical equipment and the equipment risk. The decision management module is used to provide decision support based on risk assessment results, and to help relevant organizations and personnel take risk management and control measures by generating risk reports, alerts and recommendations.

2. A data-driven medical device risk management method, applied to the data-driven medical device risk management system described in claim 1, characterized in that, Includes the following steps: S1. By collecting and storing the medical equipment itself and its operational data, a total data set is obtained, and data preprocessing is performed to reduce noise interference; S2. Based on the preprocessed medical equipment data, an anomaly detection neural network model is constructed. The anomaly detection neural network model can output the anomaly detection results of medical equipment according to the characteristics and historical data of the medical equipment; and calculate the correlation between the internal and external factors of the medical equipment and their impact on risk. S3. Conduct real-time assessments of the risks associated with the operation of medical equipment, promptly detect abnormal behavior of the equipment, and take measures based on the risk assessment results to prevent equipment malfunctions and accidents.

3. The medical device risk management method based on data analysis according to claim 2, characterized in that, Step S2 specifically includes: Where Mco represents the correlation between internal and external factors of a medical device and its risk; cov(a,b) represents the correlation between the a-th accessory and the b-th type of the medical device, where a∈[1,n], b∈[1,w]; a,1 Indicates the number of times medical equipment needs maintenance; I a,2 Indicates the number of times medical equipment needs to be repaired; β1, β2, and β3 represent the corresponding correlation parameters; O b,1 Indicates operator error; O b,2 This indicates the number of power supply defects.

4. The medical device risk management method based on data analysis according to claim 2, characterized in that, Step S2 specifically includes: A neural network model for detecting anomalies in medical devices is established using deep learning algorithms. The model includes an input layer, a detection layer, a confirmation layer, and an output layer.

5. The medical device risk management method based on data analysis according to claim 4, characterized in that, The q attribute information of the medical device data is input into the input layer and represented as follows: The input layer and the detection layer are fully connected. Anomaly detection is performed on the medical device data features in the detection layer. The specific process is as follows: INU2=ω1INU1+b1 Where INU2 represents the input in the detection layer; TOU2 represents the output of the detection layer; ω1 represents the connection weight between the input layer and the detection layer; b1 represents the bias of the detection layer; λ represents the fusion coefficient; θ represents a constant value to ensure the stability of the calculation process; α1 and α2 represent the corresponding adjustment parameters; INU 1,t This represents the average fluctuation of medical device data obtained at time t; INU 2,t α represents the standardized value of the detection layer input obtained at time t; α represents the activation function; T represents the bias; σ represents the detection parameters; and ca0 represents the preset detection parameter value.

6. The medical device risk management method based on data analysis according to claim 2, characterized in that, Step S3 specifically includes: Based on the output of the medical equipment anomaly detection neural network model and the correlation between the internal and external factors of medical equipment and their impact on the risk of medical equipment, the evaluation coefficient Rs for medical equipment risk management is calculated, defining ip internal factors and ep external factors.

7. The medical device risk management method based on data analysis according to claim 6, characterized in that, The risk of medical equipment is managed based on an assessment coefficient. Rs' is defined as a preset standard assessment parameter, τ represents the system assessment characteristic parameter, and Ran represents the risk level. If Rs' ≥ Rs... This indicates that the medical device is at high risk; if 0 ≤ Rs' <Rs, This indicates that the medical equipment is at medium risk; all other situations indicate that the medical equipment is at low risk.

Images