Infection risk assessment method and system for nursing

Abstract

The invention relates to an infection risk assessment method and system for nursing, and the method comprises the steps: collecting multi-source heterogeneous data of a patient, and fusing the multi-source heterogeneous data into a unified original data set; obtaining a standardized risk feature vector based on the original data set; inputting the risk feature vectors into a multi-modal risk assessment model, and outputting risk probabilities of individual infection and group infection; forming a personalized risk index based on the risk probability; according to the personalized risk indexes, the decision support system outputs hierarchical intervention measures matched with the risk levels, and records and feeds back execution conditions and intervention effect data of the hierarchical intervention measures for dynamically updating parameters of the risk assessment model; and integrating the intervention effect data and the early warning signal into feedback information to form a dynamic infection risk assessment system adapted to specific hospital characteristics. The accuracy of infection risk assessment of nursing can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an infection risk assessment method and system for nursing. Background Art

[0002] In clinical nursing, infection risk assessment is a crucial component in ensuring patient safety and improving nursing quality. Existing infection risk assessment methods primarily rely on nursing staff to make a preliminary assessment of a patient's infection risk during admission assessments, preoperative evaluations, or routine ward rounds. This assessment involves reviewing patient medical records, assessing underlying medical conditions, observing physical signs, and completing paper or electronic risk assessment forms.

[0003] However, assessment indicators often rely on subjective judgment and static data reporting by nursing staff, lacking the ability to comprehensively analyze dynamic data from multiple sources. Furthermore, current assessment processes are mostly manual, lacking intelligent and standardized support. This can easily lead to high-risk patients not being identified in a timely manner due to assessment omissions or delayed information updates. Summary of the Invention

[0004] The present invention addresses the technical problems existing in the prior art and provides a method and system for evaluating infection risk in nursing, which can improve the accuracy of infection risk evaluation in nursing.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: The present invention provides a method for assessing infection risk for nursing, the method comprising: Collecting multi-source heterogeneous data of patients and fusing the multi-source heterogeneous data into a unified original data set; Based on the original data set, a dynamic weight calculation algorithm is used to quantify the risk of the patient to obtain a standardized risk feature vector; Input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; Based on the risk probability, a dynamic threshold algorithm is used to score the patient's infection risk in real time, generate a risk trend curve, and trigger corresponding level warning signals according to preset multi-level thresholds to form a personalized risk index; Based on the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, and records and provides feedback on the implementation of the tiered intervention measures and intervention effect data, which are used to dynamically update the parameters of the risk assessment model; The intervention effect data and the early warning signal are integrated into feedback information, and a cross-departmental collaborative linkage mechanism is driven based on the feedback information. At the same time, the risk assessment model is continuously iterated and optimized to form a dynamic infection risk assessment system that adapts to the characteristics of specific hospitals.

[0006] Optionally, the acquisition of multi-source heterogeneous data of the patient includes: Wearable devices are used to collect patient temperature, heart rate, and blood oxygen saturation data in real time to build a dynamic physiological parameter set; Environmental sensors deployed in wards collect data on microbial concentration, temperature and humidity in the air to construct an environmental risk matrix.

[0007] Optionally, the step of quantifying the patient's risk using a dynamic weight calculation algorithm based on the original data set to obtain a standardized risk feature vector includes: Calculate the cross-exposure risk value of the current patient based on the location trajectory data of patients confirmed to be infected with MRSA in the same ward; Generate a corresponding behavioral risk correction coefficient based on the number of hand hygiene violations counted by the intelligent video recognition system; The cross-exposure risk value and the behavioral risk correction coefficient are fused and the fused result is standardized to obtain the risk feature vector.

[0008] Optionally, the risk assessment model is constructed by the following steps: performing abnormal fluctuation analysis on the patient's body temperature time series data based on a long short-term memory neural network to obtain an analysis result; Based on natural language processing, the frequency of occurrence of keywords in medical order texts is counted to generate risk labels for dysbiosis. An initial risk assessment model is trained based on the analysis results and the risk labels to obtain the risk assessment model.

[0009] Optionally, the semantic matching analysis is performed in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection, including: When it is detected that the procalcitonin concentration of the patient is greater than a preset value, matching an infection probability node associated with Gram-negative bacteria in the knowledge graph; The infection probability node is combined with the hospital's antimicrobial resistance spectrum data to output the risk probability.

[0010] Optionally, triggering a warning signal of a corresponding level according to a preset multi-level threshold includes: Real-time risk scoring is performed based on three levels of warning thresholds, where the basic threshold is a risk score greater than 7 points, the cumulative change threshold is a score increase of 2 points or more over 3 hours, and the trend threshold is a score curve slope greater than 1.0 point / hour; During the period when the patient undergoes high-risk procedures such as bronchoscopy, the frequency of risk monitoring is dynamically increased.

[0011] Optionally, forming a personalized risk index includes: When the patient is an individual who has undergone organ transplantation, the weight of the environmental microbial risk-related indicators is increased; When the patient has a history of colonization with carbapenem-resistant Enterobacteriaceae, a special risk value is added to the basic risk score.

[0012] Optionally, the decision support system outputs, based on the personalized risk index, tiered intervention measures that match the risk level, including: When the patient's risk score is greater than a first score, single-room isolation measures are automatically triggered and a high-efficiency air purification device is allocated; When the patient risk score is between a second score and the first score, a protective equipment checklist is generated and a special risk assessment task is issued n times a day; wherein the second score is less than the first score.

[0013] Optionally, the continuously iterative optimization of the risk assessment model includes: By comparing the deviation between the system prediction result and the actual infection status of the patient, the back propagation algorithm is used to optimize and adjust the parameters of the long short-term memory neural network; The algorithm for invasive procedure risk was calibrated and revised based on historical catheter-related infection data in the ICU ward.

[0014] The present invention also provides an infection risk assessment system for nursing, the system comprising: A data acquisition module is used to collect multi-source heterogeneous data of patients and fuse the multi-source heterogeneous data into a unified original data set; A risk signature module, configured to quantify the patient's risk based on the original data set using a dynamic weight calculation algorithm to obtain a standardized risk signature vector; A risk probability module is used to input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; A risk index module is used to score the patient's infection risk in real time based on the risk probability using a dynamic threshold algorithm, generate a risk trend curve, and trigger a warning signal of a corresponding level according to preset multi-level thresholds to form a personalized risk index; A parameter updating module is used to output, based on the personalized risk index, tiered intervention measures that match the risk level from the decision support system, and to record and provide feedback on the implementation of the tiered intervention measures and intervention effect data, so as to dynamically update the parameters of the risk assessment model; The risk assessment module is used to integrate the intervention effect data and the early warning signal into feedback information, drive the cross-departmental collaborative linkage mechanism based on the feedback information, and continuously iterate and optimize the risk assessment model to form a dynamic infection risk assessment system adapted to the characteristics of specific hospitals.

[0015] In addition, to achieve the above-mentioned purpose, the present invention also proposes an electronic device, comprising: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing an infection risk assessment method for nursing as described above.

[0016] In addition, to achieve the above-mentioned purpose, the present invention also proposes a non-transitory computer-readable storage medium, in which a computer software program is stored. When the computer software program is executed by a processor, it implements an infection risk assessment method for nursing as described above.

[0017] The beneficial effects of the present invention are: (1) By constructing a multi-source heterogeneous data acquisition system and a dynamic weight calculation mechanism, the present invention can capture patient status and environmental changes in real time, breaking through the limitations of traditional assessment methods that are based on static single-point data and have a low update frequency, and realizing continuous monitoring and real-time updating of infection risks.

[0018] (2) The present invention utilizes a multimodal deep learning model to integrate multiple data sources such as temporal physiological parameters, text medical order information, and knowledge graphs, significantly enhancing the model's ability to identify potential signs of infection and providing higher accuracy and response speed in early risk warning.

[0019] (3) By constructing a standardized risk feature vector and introducing a dynamic weight adjustment mechanism, the present invention effectively reduces the evaluation process's reliance on human experience and subjective judgment, and improves the consistency, repeatability, and transparency of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart of a method for infection risk assessment for nursing provided by the present invention; Figure 2This is a schematic structural diagram of an infection risk assessment system for nursing provided by the present invention; Figure 3 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention; Figure 4 A schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0023] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0024] See also Figure 1 , provides a flowchart of an infection risk assessment method for nursing of the present invention, comprising the following steps: Step 201: Collect multi-source heterogeneous data of a patient and fuse the multi-source heterogeneous data into a unified original data set.

[0025] In specific implementation, heterogeneous data from multiple dimensions such as patients, environment and nursing behavior can be integrated, such as physiological parameters, electronic medical records, test indicators, ward environment sensor data and medical operation records, to build a structurally unified and semantically consistent original data set, providing a comprehensive foundation for subsequent risk modeling.

[0026] In some embodiments, step 201 may include: Wearable devices are used to collect patient temperature, heart rate, and blood oxygen saturation data in real time to build a dynamic physiological parameter set; Environmental sensors deployed in wards collect data on microbial concentration, temperature and humidity in the air to construct an environmental risk matrix.

[0027] In a preferred embodiment of the present invention, the collection of multi-source heterogeneous data includes the acquisition and structured processing of patient physiological parameter data and hospital environment data, specifically including the following: First, the system connects to wearable physiological monitoring devices worn by patients to continuously collect core vital signs in real time, including but not limited to temperature, heart rate, and blood oxygen saturation. These physiological parameter data are timestamped and standardized to construct a dynamic physiological parameter set that reflects the patient's current and recent health status trends.

[0028] Secondly, air quality monitoring sensors are deployed in the patient's ward environment to continuously collect environmental risk factors, including airborne microbial concentration, temperature, and humidity. By spatially calibrating and thresholding this environmental monitoring data, an environmental risk matrix with a time dimension and risk score attributes is generated. This matrix can reflect the potential risk level of infection transmission in a specific ward during a specific time period.

[0029] Through the joint collection and structured expression of the above two types of physiological and environmental data, a multi-dimensional and dynamic data foundation is provided for subsequent risk feature extraction, infection probability prediction and personalized intervention recommendations.

[0030] Step 202: Based on the original data set, a dynamic weight calculation algorithm is used to quantify the patient's risk to obtain a standardized risk feature vector.

[0031] In specific implementation, a dynamic weight calculation algorithm can be used to process the original data, assign weights based on the importance and time sensitivity of different risk factors, and then standardize and structure various risk characteristics to form a unified risk feature vector for model analysis.

[0032] In some embodiments, step 202 may include: Calculate the cross-exposure risk value of the current patient based on the location trajectory data of patients confirmed to be infected with MRSA in the same ward; Generate a corresponding behavioral risk correction coefficient based on the number of hand hygiene violations counted by the intelligent video recognition system; The cross-exposure risk value and the behavioral risk correction coefficient are fused and the fused result is standardized to obtain the risk feature vector.

[0033] In a preferred embodiment of the present invention, the construction of the risk feature vector includes extracting and fusing the cross-exposure risk value and the behavioral risk correction coefficient, specifically including the following steps: First, the system accesses the ward's patient location system to obtain the historical activity trajectories of patients diagnosed with methicillin-resistant Staphylococcus aureus (MRSA) infection in the same ward. Based on the spatial-temporal overlap between this location trajectory and the currently assessed patient, combined with factors such as contact duration and proximity, a spatial intersection modeling and weighted integral algorithm are used to calculate the current patient's cross-exposure risk value. This value is used to measure the potential infection risk faced by the patient due to spatial contact.

[0034] Secondly, the system automatically detects and counts the hand hygiene behaviors of medical staff during daily operations through intelligent video recognition devices deployed in wards. Based on preset hand hygiene behavior standards, the system identifies and counts hand hygiene violations. After normalization, the violation data is mapped to a quantitative behavioral risk correction factor, which is used to correct the patient's passive exposure risk during nursing operations.

[0035] Furthermore, the system fuses the above-mentioned cross-exposure risk value with the behavioral risk correction coefficient, and uses a standardization function to normalize the units and scales of the fusion results to obtain a structured risk feature vector, which is subsequently input into the multimodal deep learning model to participate in the prediction calculation of infection risk.

[0036] This implementation method significantly enhances the responsiveness of risk modeling to ward transmission paths and nursing behavior compliance by introducing two behavioral correlation factors: spatial contact and operational norms, and improves the clinical relevance and credibility of the evaluation results.

[0037] Step 203: Input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection.

[0038] In specific implementation, the standardized risk feature vector can be input into the multimodal risk assessment model. The model predicts the infection probability of individuals and groups by learning time series features and deep association modeling, combined with the clinical-pathogen semantic relationship in the infectious microorganism knowledge graph.

[0039] In some embodiments, the risk assessment model may be constructed by the following steps: performing abnormal fluctuation analysis on the patient's body temperature time series data based on a long short-term memory neural network to obtain an analysis result; Based on natural language processing, the frequency of occurrence of keywords in medical order texts is counted to generate risk labels for dysbiosis. An initial risk assessment model is trained based on the analysis results and the risk labels to obtain the risk assessment model.

[0040] In a preferred embodiment of the present invention, the construction of the risk assessment model includes two stages: multimodal feature learning and fusion modeling, specifically including the following steps: First, the system models the patient's temperature time series data using a long short-term memory (LSTM) neural network. This LSTM network is capable of learning long-term dependencies within sequence data, enabling in-depth analysis of potential trends, cyclical fluctuations, and abnormal increases in temperature data. By analyzing abnormal fluctuations in temperature time series data over timescales of 72 hours or longer, the system outputs analysis results that characterize the patient's recent infection risk trends.

[0041] Secondly, the system performs semantic analysis on the patient's medical order text using natural language processing (NLP). Specifically, the system constructs a keyword library for infection risk, consisting of keywords such as "antibiotic category," "dose," and "frequency of use." The system then counts the frequency of these keywords within the medical order text and performs contextual semantic matching to determine whether there is abuse or reuse of broad-spectrum antibiotics. Based on these results, a risk tag is generated to represent the patient's potential for dysbiosis, identifying potential infection risk factors associated with microbial imbalance.

[0042] Furthermore, the system integrates the LSTM analysis results with the risk labels generated by natural language processing, employing a supervised learning strategy to train an initial multimodal infection risk assessment model. This model integrates information from both time series and text semantics to achieve a probabilistic prediction of a patient's future infection risk, thereby forming a risk assessment model with structured input and generalizability.

[0043] The risk assessment model constructed in the above way has both sensitivity to the dynamic changes of physiological signals and the ability to extract potential risk factors in semantic information, providing key support for subsequent personalized infection risk prediction and dynamic early warning.

[0044] In some embodiments, step 203 may include: When it is detected that the procalcitonin concentration of the patient is greater than a preset value, matching an infection probability node associated with Gram-negative bacteria in the knowledge graph; The infection probability node is combined with the hospital's antimicrobial resistance spectrum data to output the risk probability.

[0045] In a preferred embodiment of the present invention, the output process of the infection risk probability includes knowledge graph matching analysis and drug resistance data fusion calculation, specifically including the following steps: First, while the system continuously monitors patient test data, if it detects a procalcitonin (PCT) concentration exceeding a preset threshold (e.g., 2 ng / mL), it triggers a semantic matching mechanism related to infection. This mechanism calls upon a pre-built knowledge graph of infectious microorganisms, which represents the relationships between common infectious pathogens, pathogenic mechanisms, biomarkers, and clinical manifestations in the form of a graph database.

[0046] The system uses "procalcitonin > threshold" as the query condition in the graph, performs backward semantic expansion operations, automatically matches probability nodes related to infection with Gram-negative bacteria (such as Pseudomonas aeruginosa, Escherichia coli, etc.), and calculates the preliminary probability score of the current patient being infected with this type of pathogen based on factors such as the edge weights between nodes in the graph, clinical relevance, and literature support.

[0047] The system further calibrates these preliminary probability scores based on the hospital's antimicrobial resistance spectrum data. This data, generated from historical bacterial culture and susceptibility testing results within the hospital, reflects the resistance trends of specific bacterial species to commonly used antibiotics in specific wards.

[0048] The system cross-calculates the Gram-negative bacterial species that the patient may be infected with with the drug resistance data of the corresponding bacterial species in similar scenarios in this hospital, dynamically adjusts the infection possibility score, and finally outputs the infection risk probability that integrates the microbial knowledge graph and local drug resistance characteristics as one of the input variables of the risk assessment model.

[0049] Through the above mechanism, the system not only realizes the intelligent identification of the infection type indicated by elevated PCT, but also improves the accuracy and adaptability of prediction by combining localized clinical data, effectively enhancing the interpretability and clinical guidance value of the risk assessment model.

[0050] Step 204: Based on the risk probability, a dynamic threshold algorithm is used to score the patient's infection risk in real time, generate a risk trend curve, and trigger a warning signal of a corresponding level according to preset multi-level thresholds to form a personalized risk index; In practice, a dynamic threshold strategy can be applied to the infection risk probability output by the model to update the risk score in real time, generating a patient risk trend curve. Based on the magnitude, trend, and absolute value of the risk value, a warning signal of the corresponding level is automatically triggered to construct a personalized infection risk index.

[0051] In some embodiments, step 204 may include: Real-time risk scoring is performed based on three levels of warning thresholds, where the basic threshold is a risk score greater than 7 points, the cumulative change threshold is a score increase of 2 points or more over 3 hours, and the trend threshold is a score curve slope greater than 1.0 point / hour; During the period when the patient undergoes high-risk procedures such as bronchoscopy, the frequency of risk monitoring is dynamically increased.

[0052] In a preferred embodiment of the present invention, to achieve continuous assessment and graded early warning of infection risk, the system establishes a three-level early warning scoring mechanism based on a set multi-level dynamic threshold strategy, and intelligently adjusts the risk monitoring frequency in combination with clinical high-risk operation events, specifically including the following: First, the system calculates the patient's personalized infection risk index in real time and dynamically scores and makes early warning judgments based on the following three thresholds: Basic threshold: When the current infection risk score is greater than 7 points, the system identifies it as a potential high-risk state and automatically triggers a level 1 warning; Cumulative change threshold: When the patient's infection risk score increases by 2 points or more within 3 consecutive hours, it is judged as a rapid increase in risk within a short period of time, triggering a second-level warning; Trend threshold: When the growth slope of the risk score curve exceeds 1.0 point / hour, it indicates that the risk is showing a continuously accelerating upward trend, and the system triggers a Level 3 trend warning.

[0053] The above three dimensions together constitute a multi-dimensional risk change identification mechanism. The system can make comprehensive judgments based on the current score value, short-term fluctuations and score growth rate to improve the response speed and sensitivity to potential infection events.

[0054] Furthermore, the system is equipped with a clinical context sensitivity adjustment module to identify whether the patient is currently undergoing a high-risk infection procedure, such as a bronchoscopy, intubation, tracheotomy, or other invasive procedure. If it detects that such a high-risk procedure is in progress or about to begin, the system automatically increases the infection risk score update frequency from the default period (e.g., every 30 minutes) to a higher frequency (e.g., every 10 minutes) to capture subtle changes in the patient's physiological state and infection risk before and after the procedure in real time.

[0055] Through the above-mentioned scoring threshold determination and dynamic adjustment mechanism of monitoring frequency, the present invention can achieve real-time perception, rapid response and hierarchical control of infection risks, provide reliable triggering basis and time window guarantee for subsequent intervention strategies, and enhance the intelligence and clinical adaptability of the overall infection early warning system.

[0056] In some embodiments, step 204 may further include: When the patient is an individual who has undergone organ transplantation, the weight of the environmental microbial risk-related indicators is increased; When the patient has a history of colonization with carbapenem-resistant Enterobacteriaceae, a special risk value is added to the basic risk score.

[0057] In a preferred embodiment of the present invention, to achieve personalized and accurate risk assessment, the system introduces a patient characteristic sensitivity adjustment mechanism during the risk feature vector generation and risk score calculation process, dynamically weighting the relevant risk factors of specific high-risk patient groups, specifically including the following: First, when the system identifies a patient as a post-organ transplant recipient, it automatically applies a risk-weighting strategy based on their immunosuppressive status. Because these patients typically undergo long-term immunosuppressive therapy, their resistance to environmental pathogens is significantly reduced, leading to a significantly increased susceptibility to infection. Based on this clinical characteristic, the system assigns enhanced weighting coefficients to environmental microbial indicators (such as ward air microbial concentration and surface contamination risk level) that comprise the risk feature vector. For example, the weighting may be increased to 150% of the baseline value to reflect the increased pathogenic potential of these environmental exposures.

[0058] Secondly, if a patient's medical record indicates a history of colonization with carbapenem-resistant Enterobacteriaceae (CRE), the system automatically loads a specialized risk assessment module related to drug resistance. Considering that this type of colonization can persist for a long time and, under certain conditions, transform into a clinical infection, the system adds an additional specialized risk value to the basic infection risk score, such as a fixed weighted item of 3.0 points, to enhance the model's sensitivity to this specific historical factor.

[0059] By adjusting the weights and correcting the scores of the above two types of individualized risk factors, the system can accurately express risks based on the patient's unique immune status and previous history of infection, significantly improving the adaptability, pertinence and clinical practicality of the risk assessment model, and ensuring the identification efficiency and timeliness of intervention in high-risk populations.

[0060] Step 205: Based on the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, and records and provides feedback on the implementation status and intervention effect data of the tiered intervention measures to dynamically update the parameters of the risk assessment model.

[0061] In a specific implementation, the system can call a preset intervention knowledge base based on the personalized risk index and output hierarchical intervention measures corresponding to the risk level; and record the execution process of the measures and their clinical effects in real time to evaluate the effectiveness of the intervention and assist in model self-optimization.

[0062] In some embodiments, step 205 may include: When the patient's risk score is greater than a first score, single-room isolation measures are automatically triggered and a high-efficiency air purification device is allocated; When the patient risk score is between a second score and the first score, a protective equipment checklist is generated and a special risk assessment task is issued n times a day; wherein the second score is less than the first score.

[0063] In a preferred embodiment of the present invention, to achieve hierarchical response and precise control of infection risk intervention, the system sets multiple threshold segments based on the patient's real-time infection risk score and automatically links different levels of prevention and control measures, specifically including the following: When the system detects that the patient's current infection risk score is greater than the preset first score (for example, the first score is 9 points), the patient is determined to be in a high-risk range and the system automatically triggers high-level infection prevention and control measures, including: Send instructions to the ward management system to transfer or arrange the patient to a single ward with physical isolation conditions; At the same time, air purification devices with high-efficiency particulate matter filtration function (such as HEPA-level air filtration system) are deployed in the patient's ward or treatment area to reduce the exposure concentration of airborne infectious factors.

[0064] If the patient's risk score is between the second score and the first score (i.e., the second score < risk score ≤ first score, for example, the second score is 6), the system determines that the patient is a medium-risk individual and automatically generates corresponding intermediate intervention measures, including: Output a protective equipment checklist containing core supplies such as hand hygiene products, masks, and protective clothing, and push it to the mobile terminals of responsible nurses and infection control personnel to ensure that medical protection standards are implemented without omissions; Based on the changes in the risk trend curve, special infection risk assessment tasks are set and issued n times a day (for example, 3 times), including bedside rounds of patients, retesting of physiological indicators, supervision of nursing operation processes, etc., to ensure dynamic grasp and response to risk changes.

[0065] Through the setting and systematic linkage of the above-mentioned hierarchical intervention strategy, the present invention not only improves the response efficiency and refinement of infection prevention and control measures, but also effectively reduces the waste of protective resources and intervention blind spots, achieves accurate matching of risk level and intervention intensity, and enhances the closed-loop management capability of nursing infection control throughout the entire process.

[0066] Step 206: Integrate the intervention effect data and the early warning signal into feedback information, drive a cross-departmental collaborative linkage mechanism based on the feedback information, and continuously iterate and optimize the risk assessment model to form a dynamic infection risk assessment system that adapts to the characteristics of a specific hospital.

[0067] In specific implementation, intervention effect data and early warning signals can be integrated to generate feedback information, and promote information linkage and resource coordination among relevant departments such as infection control, nursing, and pharmacy; at the same time, feedback data can be used to continuously optimize model parameters and rule settings, thereby building an adaptive dynamic risk assessment system for specific medical institution scenarios.

[0068] In some embodiments, step 206 may include: By comparing the deviation between the system prediction result and the actual infection status of the patient, the back propagation algorithm is used to optimize and adjust the parameters of the long short-term memory neural network; The algorithm for invasive procedure risk was calibrated and revised based on historical catheter-related infection data in the ICU ward.

[0069] In a preferred embodiment of the present invention, in order to achieve continuous optimization and hospital-specific adaptation of the infection risk assessment model, the system constructs a model self-learning and characteristic factor calibration mechanism based on feedback data, which specifically includes the following steps: First, the system predicts each patient's infection risk during the evaluation period and generates an infection probability value or risk level label. This prediction is then compared with the patient's actual infection history during clinical practice, and the deviation between the two is calculated. This deviation can be measured using an error function (such as cross-entropy loss or mean square error loss).

[0070] Based on these deviations, the system invokes a backpropagation algorithm to perform a gradient update on the weight parameters in the long-short-term memory (LSTM) neural network model used to construct infection risk trends. Through multiple rounds of error feedback and parameter adjustments, the model continuously adapts to the characteristics of current in-hospital patient data, thereby improving prediction accuracy and generalization capabilities.

[0071] Furthermore, the system combines historical data on catheter-related infection events in the intensive care unit (ICU), including infection probability, average incubation period, and intervention effectiveness for procedures such as urinary catheters, central venous catheters, and tracheostomies, to locally calibrate the algorithm parameters currently used to calculate invasive procedure risk factors. For example, variables such as catheter indwelling duration, insertion site type, and procedural compliance in the original risk scoring model can be assigned updated weighting coefficients to make the risk expression more consistent with the sensitivity and time-dependence of infection risk in actual ICU scenarios.

[0072] Through the above-mentioned prediction bias feedback optimization and historical data-driven calibration mechanism, the infection risk assessment system constructed by the present invention can achieve adaptive evolution based on empirical data, enhancing its applicability and clinical reference value in different wards, different patient groups and different operation scenarios.

[0073] See also Figure 2 , Figure 2 This is a schematic structural diagram of an infection risk assessment system for nursing provided by the present invention.

[0074] like Figure 2 As shown, an infection risk assessment system for nursing proposed in an embodiment of the present invention includes: The data acquisition module 301 is used to collect multi-source heterogeneous data of patients and fuse the multi-source heterogeneous data into a unified original data set; A risk signature module 302 is configured to quantify the patient's risk using a dynamic weight calculation algorithm based on the original data set to obtain a standardized risk signature vector; The risk probability module 303 is used to input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; The risk index module 304 is configured to use a dynamic threshold algorithm to score the patient's infection risk in real time based on the risk probability, generate a risk trend curve, and trigger a warning signal of a corresponding level according to preset multi-level thresholds to form a personalized risk index; Parameter updating module 305 is used to output, based on the personalized risk index, tiered intervention measures that match the risk level from the decision support system, and record and provide feedback on the implementation of the tiered intervention measures and intervention effect data, so as to dynamically update the parameters of the risk assessment model; The risk assessment module 306 is used to integrate the intervention effect data and the early warning signal into feedback information, drive the cross-departmental collaborative linkage mechanism based on the feedback information, and continuously iterate and optimize the risk assessment model to form a dynamic infection risk assessment system that adapts to the characteristics of a specific hospital.

[0075] See also Figure 3 , Figure 3 Schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented: Collecting multi-source heterogeneous data of patients and fusing the multi-source heterogeneous data into a unified original data set; Based on the original data set, a dynamic weight calculation algorithm is used to quantify the risk of the patient to obtain a standardized risk feature vector; Input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; Based on the risk probability, a dynamic threshold algorithm is used to score the patient's infection risk in real time, generate a risk trend curve, and trigger corresponding level warning signals according to preset multi-level thresholds to form a personalized risk index; Based on the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, and records and provides feedback on the implementation of the tiered intervention measures and intervention effect data, which are used to dynamically update the parameters of the risk assessment model; The intervention effect data and the early warning signal are integrated into feedback information, and a cross-departmental collaborative linkage mechanism is driven based on the feedback information. At the same time, the risk assessment model is continuously iterated and optimized to form a dynamic infection risk assessment system that adapts to the characteristics of specific hospitals.

[0076] See also Figure 4 , Figure 4 Schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4As shown, this embodiment provides a computer-readable storage medium 500 on which a computer program 411 is stored. When the computer program 411 is executed by a processor, the following steps are implemented: Collecting multi-source heterogeneous data of patients and fusing the multi-source heterogeneous data into a unified original data set; Based on the original data set, a dynamic weight calculation algorithm is used to quantify the risk of the patient to obtain a standardized risk feature vector; Input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; Based on the risk probability, a dynamic threshold algorithm is used to score the patient's infection risk in real time, generate a risk trend curve, and trigger corresponding level warning signals according to preset multi-level thresholds to form a personalized risk index; Based on the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, and records and provides feedback on the implementation of the tiered intervention measures and intervention effect data, which are used to dynamically update the parameters of the risk assessment model; The intervention effect data and the early warning signal are integrated into feedback information, and a cross-departmental collaborative linkage mechanism is driven based on the feedback information. At the same time, the risk assessment model is continuously iterated and optimized to form a dynamic infection risk assessment system that adapts to the characteristics of specific hospitals.

[0077] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0078] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0079] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or boxes.

[0080] These computer program instructions may 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 produce an article of manufacture including an instruction system that is implemented in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0082] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0083] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for infection risk assessment for nursing, characterized in that: The method comprises: Collecting multi-source heterogeneous data of patients and fusing the multi-source heterogeneous data into a unified original data set; Based on the original data set, a dynamic weight calculation algorithm is used to quantify the risk of the patient to obtain a standardized risk feature vector; Input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; Based on the risk probability, a dynamic threshold algorithm is used to score the patient's infection risk in real time, generate a risk trend curve, and trigger corresponding level warning signals according to preset multi-level thresholds to form a personalized risk index; Based on the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, and records and provides feedback on the implementation of the tiered intervention measures and intervention effect data, which are used to dynamically update the parameters of the risk assessment model; The intervention effect data and the early warning signal are integrated into feedback information, and a cross-departmental collaborative linkage mechanism is driven based on the feedback information. At the same time, the risk assessment model is continuously iterated and optimized to form a dynamic infection risk assessment system that adapts to the characteristics of specific hospitals.

2. The method according to claim 1, characterized in that The multi-source heterogeneous data of the patient is collected, including: Wearable devices are used to collect patient temperature, heart rate, and blood oxygen saturation data in real time to build a dynamic physiological parameter set; Environmental sensors deployed in wards collect data on microbial concentration, temperature and humidity in the air to construct an environmental risk matrix.

3. The method according to claim 2, characterized in that The method of quantifying the patient's risk based on the original data set using a dynamic weight calculation algorithm to obtain a standardized risk feature vector includes: Calculate the cross-exposure risk value of the current patient based on the location trajectory data of patients confirmed to be infected with MRSA in the same ward; Generate a corresponding behavioral risk correction coefficient based on the number of hand hygiene violations counted by the intelligent video recognition system; The cross-exposure risk value and the behavioral risk correction coefficient are fused and the fused result is standardized to obtain the risk feature vector.

4. The method according to claim 3, characterized in that The risk assessment model is constructed by the following steps: performing abnormal fluctuation analysis on the patient's body temperature time series data based on a long short-term memory neural network to obtain an analysis result; Based on natural language processing, the frequency of occurrence of keywords in medical order texts is counted to generate risk labels for dysbiosis. An initial risk assessment model is trained based on the analysis results and the risk labels to obtain the risk assessment model.

5. The method according to claim 4, characterized in that The semantic matching analysis is performed in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection, including: When it is detected that the procalcitonin concentration of the patient is greater than a preset value, matching an infection probability node associated with Gram-negative bacteria in the knowledge graph; The infection probability node is combined with the hospital's antimicrobial resistance spectrum data to output the risk probability.

6. The method according to claim 5, characterized in that The triggering of the corresponding level of warning signal according to the preset multi-level threshold includes: Real-time risk scoring is performed based on three levels of warning thresholds, where the basic threshold is a risk score greater than 7 points, the cumulative change threshold is a score increase of 2 points or more over 3 hours, and the trend threshold is a score curve slope greater than 1.0 point / hour; During the period when the patient undergoes high-risk procedures such as bronchoscopy, the frequency of risk monitoring is dynamically increased.

7. The method according to claim 6, characterized in that The forming of a personalized risk index includes: When the patient is an individual who has undergone organ transplantation, the weight of the environmental microbial risk-related indicators is increased; When the patient has a history of colonization with carbapenem-resistant Enterobacteriaceae, a special risk value is added to the basic risk score.

8. The method according to claim 7, characterized in that According to the personalized risk index, the decision support system outputs tiered intervention measures that match the risk level, including: When the patient's risk score is greater than a first score, single-room isolation measures are automatically triggered and a high-efficiency air purification device is allocated; When the patient risk score is between a second score and the first score, a protective equipment checklist is generated and a special risk assessment task is issued n times a day; wherein the second score is less than the first score.

9. The method according to claim 8, characterized in that The continuously iterative optimization of the risk assessment model includes: By comparing the deviation between the system prediction result and the actual infection status of the patient, the back propagation algorithm is used to optimize and adjust the parameters of the long short-term memory neural network; The algorithm for invasive procedure risk was calibrated and revised based on historical catheter-related infection data in the ICU ward.

10. An infection risk assessment system for nursing, characterized in that: The system comprises: A data acquisition module is used to collect multi-source heterogeneous data of patients and fuse the multi-source heterogeneous data into a unified original data set; A risk signature module, configured to quantify the patient's risk based on the original data set using a dynamic weight calculation algorithm to obtain a standardized risk signature vector; A risk probability module is used to input the risk feature vector into a multimodal risk assessment model, use the risk assessment model to perform feature learning on the patient's time series data, and perform semantic matching analysis in combination with the infectious microorganism knowledge graph to output the risk probability of individual infection and group infection; A risk index module is used to score the patient's infection risk in real time based on the risk probability using a dynamic threshold algorithm, generate a risk trend curve, and trigger a warning signal of a corresponding level according to preset multi-level thresholds to form a personalized risk index; A parameter updating module is used to output, based on the personalized risk index, tiered intervention measures that match the risk level from the decision support system, and to record and provide feedback on the implementation of the tiered intervention measures and intervention effect data, so as to dynamically update the parameters of the risk assessment model; The risk assessment module is used to integrate the intervention effect data and the early warning signal into feedback information, drive the cross-departmental collaborative linkage mechanism based on the feedback information, and continuously iterate and optimize the risk assessment model to form a dynamic infection risk assessment system adapted to the characteristics of specific hospitals.

Images