Patient safety management informatization method and system based on medical quality control model

Abstract

The invention discloses a patient safety management informatization method and system based on a medical quality control model, and relates to the technical field of modern medical treatment, and the method comprises the steps: obtaining patient condition data based on a hospital information system; the method comprises the following steps: collecting data of a patient, training the collected data based on a dual-machine learning algorithm in combination with a boosting method, establishing a high-precision risk prediction model, and based on calculation of the risk prediction model, outputting complication, complication risk and death risk prediction of the patient; summarizing model prediction results to form a comprehensive risk assessment report; and early warning information is pushed to medical personnel in real time to assist in early intervention. Through a double-machine learning algorithm and a boosting method, the risk prediction precision is improved, medical negligence is reduced, the system can generate personalized early warning reports in real time, assist doctors in intervention and reduce complications, complication and death risks, through comprehensive risk assessment and personalized suggestions, medical resource allocation is optimized, the treatment effect is improved, and the risk prediction efficiency is improved. And medical misdiagnosis is reduced.

Description

Technical Field

[0001] The present invention relates to the field of modern medical technology, and in particular to a patient safety management informationization method and system based on a medical quality control model. Background Art

[0002] With the development of modern medical technology and informatization, the demand for hospital management, patient health management, and safety management is increasing. The medical charge model is a risk prediction model used in the medical field in recent years. By analyzing patients' medical records, treatment records, medication information, and other health indicators, it can predict the risks of complications, comorbidities, and death that patients may face. However, there are currently some common problems in medical safety management in the medical industry, such as incomplete patient risk assessment, inaccurate intervention timing, and uneven distribution of medical resources. Therefore, how to improve the accuracy and timeliness of patient safety management through intelligent means has become an important topic in the current construction of medical informatization.

[0003] The present invention's patient safety management informationization method and system, based on a medical quality control model, aims to utilize advanced machine learning techniques, combined with patient condition data from hospital information systems, to comprehensively analyze patients' health status, provide early warnings of potential medical risks, and assist physicians in making accurate decisions, thereby improving patient safety management effectiveness. In particular, the combination of dual machine learning algorithms and boosting methods not only improves the accuracy of risk prediction models, but also effectively reduces misdiagnosis rates and provides medical personnel with real-time early warning information, enabling timely intervention. Summary of the Invention

[0004] In order to solve the above technical problems, a patient safety management informationization method and system based on the medical quality control model are provided. This technical solution solves the above problems.

[0005] In order to achieve the above objects, the technical solution adopted by the present invention is: Patient safety management informationization methods based on medical quality control models include: Obtain patient condition data based on hospital information systems; The collected data is trained based on a dual machine learning algorithm combined with a boosting method to establish a high-precision risk prediction model. Based on the calculations of the risk prediction model, the patient's risk of complications, comorbidities, and mortality is predicted. The model prediction results are summarized to form a comprehensive risk assessment report. Push early warning information to medical personnel in real time to assist in early intervention.

[0006] Preferably, the dual machine learning algorithm is combined with the boosting method to train the collected data to establish a high-precision risk prediction model, and the output of the patient's complication, comorbidity risk and death risk prediction based on the calculation of the risk prediction model specifically includes: Extract patient medical records from the hospital information system, including historical medical conditions, physical examination data, treatment records, medication information, and unsafe incident data, and clean, denoise, and standardize the raw data; Extract key features from medical record data, including patient age, gender, past medical history, current condition, and medication use, and construct these features into a model input format; Divide the preprocessed data into training set, validation set and test set; It uses two machine learning algorithms, decision tree and support vector machine, combined with boosting method, to iterate weighted training step by step, gradually correct the errors of the previous model. During the training process, cross-validation is used to optimize the model's hyperparameters. Apply the trained model to the medical records of new patients and calculate the risk of complications, comorbidities, and death based on the feature input; Make independent predictions for different types of risks and generate probability values ​​for each risk; Summarize the prediction results of complication, comorbidity and mortality risk models, combine the predicted probabilities of different risks, and form a comprehensive risk assessment report; Based on the risk assessment results, a personalized early warning report is generated for the patient, pushed to medical staff in real time, and pushed to the doctor's workstation through a software plug-in in the hospital information system. Based on the severity of the risk, high-risk patients are given priority to assist doctors in early intervention.

[0007] Preferably, the two machine learning algorithms of decision tree and support vector machine are used in combination with boosting method, and weighted training is performed iteratively step by step to correct the error of the previous model. During the training process, the hyperparameters of the model are optimized by cross-validation, specifically including: The model formula is: f(X) represents the model's predicted output for the input feature vector X, t represents the number of steps in the current iteration, and α t represents the weight coefficient of the t-th weak learner, h t (X) represents the predicted output of the t-th weak learner for the input data X, T represents the number of iterations in the Boosting algorithm, P(R i |X) represents the i-th risk. When the input feature vector X is given, the predicted probability of the model output is f i(X) represents the prediction function of risk prediction model i, X represents the input feature vector, and i represents the index of risk type.

[0008] Preferably, the summary of the prediction results of the complication, comorbidity and mortality risk models, combining the predicted probabilities of different risks to form a comprehensive risk assessment report specifically includes: Based on Min-Max normalization, each probability value is normalized to between 0 and 1; The predicted probabilities of different risks are combined using a weighted average method to obtain a comprehensive risk score, assigning a weight to each risk; Based on the comprehensive risk score, thresholds are set to determine the patient's risk level, which is divided into low, medium and high levels. Based on each patient's comprehensive risk score and classification, a comprehensive risk assessment report is generated. The report content includes the patient's basic information, the individual prediction results of each risk model, the comprehensive risk score, and recommended measures for risk assessment.

[0009] Preferably, the weighted average method is used to combine the predicted probabilities of different risks to obtain a comprehensive risk score, and the weight allocation for each risk specifically includes: The comprehensive risk score calculation formula is: R T =l1·P(R1|X)+l2·P(R2|X)+l3·P(R3|X)+l i ·P(R i |X) Where R T is the comprehensive risk score, P(R1|X) is the predicted probability of complication risk, l1 is the precursor of the predicted probability of complication risk, P(R2|X) is the predicted probability of complication risk, l2 is the weight of the predicted probability of complication risk, P(R3|X) is the predicted probability of death risk, l3 is the weight of the predicted probability of death risk, P(R i |X) is the probability of the i-th risk prediction, l i is the probability weight of the i-th risk prediction.

[0010] Preferably, the risk assessment results are used to generate personalized early warning reports for patients, which are pushed to medical personnel in real time and to doctors' workstations via software plug-ins in the hospital information system. Based on the severity of the risk, high-risk patients are given priority, and doctors are assisted in early intervention. Specifically, the following are included: Generate a personalized early warning report based on the assessment results, which includes the patient's risk type, the predicted probability of each risk and its risk level, recommended intervention measures, and priority classification; Through the real-time data push function within the hospital information system, the generated early warning report will be sent to relevant medical personnel in real time and pushed to the doctor's workstation through the software plug-in; Based on early warning reports, the system can provide decision-making support, automatically recommend intervention measures for high-risk patients, and provide best practice guidelines for early intervention; The system continuously tracks the patient's health status and updates risk assessments in real time.

[0011] Patient safety management information system based on medical quality control model, including: Data collection module, used to automatically capture patient medical history and unsafe incident data from the hospital information system; The risk assessment module uses a dual machine learning model to process and analyze the collected data to generate a prediction of the patient's risk of complications, comorbidities, and mortality; The result integration and report generation module is used to summarize the prediction results of each risk assessment model and generate a comprehensive risk assessment report; the early warning push module is used to push early warning information to the medical staff's workstation in real time for early intervention.

[0012] Preferably, the risk assessment module specifically includes: Data pre-processing unit: Extracts patient medical records from the hospital information system, including historical conditions, physical examination data, treatment records, medication information, and unsafe event data. It then cleans, denoises, and standardizes the raw data to ensure data quality and consistency. Feature extraction unit: extracts key features and constructs these features into the format of model input; Data partitioning unit: divides the preprocessed data into training set, validation set and test set; Model training unit: uses decision tree and support vector machine machine learning algorithms, combined with boosting methods for model training, gradual weighted training, corrects the errors of the previous model, and optimizes the model's hyperparameters through cross-validation; Risk prediction unit: Based on the trained model, it predicts the medical records of new patients, outputs the prediction results of complications, comorbidities and death risks, makes independent predictions for different types of risks, and generates probability values ​​for each risk.

[0013] Preferably, the result integration and report generation module specifically includes: Risk Assessment Summary Unit: Summarizes the prediction results of complications, comorbidities and mortality risk models into a comprehensive risk assessment report; Data normalization unit: Use the Min-Max normalization method to normalize the probability value of each risk prediction to between 0 and 1; Weighted average calculation unit: performs weighted average calculation on the predicted probabilities of different risks, assigns weights to each risk according to its importance, and obtains a comprehensive risk score; Risk level classification unit: Based on the comprehensive risk score, set thresholds to determine the patient's risk level and classify it; Report generation unit: Combines the patient's basic information, the prediction results of each risk model, the comprehensive risk score and the recommended intervention measures to generate a complete comprehensive risk assessment report.

[0014] Preferably, the warning push module includes a unit Real-time data push unit: Based on the patient's risk assessment results, personalized early warning reports are pushed to medical staff in real time. The push function within the hospital information system is used to transmit the reports directly to the doctor's workstation; Risk Priority Classification Unit: Prioritizes patients based on their comprehensive risk scores to ensure that high-risk patients receive priority treatment; Decision-making support unit: Based on risk assessment reports, it provides doctors with decision-making support, automatically recommends intervention measures, and helps doctors conduct early intervention; Health status monitoring unit: The system continuously tracks the patient's health status and updates risk assessments in real time.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This invention proposes to improve the accuracy of risk prediction and reduce medical accidents through the use of dual machine learning algorithms and boosting methods. The system can generate personalized early warning reports in real time to assist doctors in intervention, reduce the risks of complications, comorbidities and death, optimize the allocation of medical resources through comprehensive risk assessment and personalized recommendations, improve treatment effects, reduce medical misdiagnosis, support intelligent decision-making, provide real-time support, improve medical management, and enhance patient safety and medical efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a step flow chart of the present invention; Figure 2 This is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0017] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are merely examples, and those skilled in the art may conceive of other obvious variations.

[0018] Reference Figure 1 As shown in Figure 1, the patient safety management informationization method based on the medical quality control model includes: Obtain patient condition data based on hospital information systems; The collected data is trained based on a dual machine learning algorithm combined with a boosting method to establish a high-precision risk prediction model. Based on the calculations of the risk prediction model, the patient's risk of complications, comorbidities, and mortality is predicted. The model prediction results are summarized to form a comprehensive risk assessment report. Push early warning information to medical personnel in real time to assist in early intervention.

[0019] The collected data is trained based on a dual machine learning algorithm combined with the boosting method to establish a high-precision risk prediction model. Based on the calculation of the risk prediction model, the patient's complication, comorbidity and mortality risk are predicted and output, including: Extract patient medical records from the hospital information system, including historical medical conditions, physical examination data, treatment records, medication information, and unsafe incident data, and clean, denoise, and standardize the raw data; Extract key features from medical record data, including patient age, gender, past medical history, current condition, and medication use, and construct these features into a model input format; Divide the preprocessed data into training set, validation set and test set; It uses two machine learning algorithms, decision tree and support vector machine, combined with boosting method, to iterate weighted training step by step, gradually correct the errors of the previous model. During the training process, cross-validation is used to optimize the model's hyperparameters. Apply the trained model to the medical records of new patients and calculate the risk of complications, comorbidities, and death based on the feature input; Make independent predictions for different types of risks and generate probability values ​​for each risk; Summarize the prediction results of complication, comorbidity and mortality risk models, combine the predicted probabilities of different risks, and form a comprehensive risk assessment report; Based on the risk assessment results, a personalized early warning report is generated for each patient and pushed to medical staff in real time. It is also pushed to the doctor's workstation via a software plug-in in the hospital information system. Based on the severity of the risk, high-risk patients are prioritized to assist doctors in early intervention. The combination of decision tree and support vector machine algorithms fully utilizes the advantages of the two algorithms, especially in the modeling of nonlinear data. The boosting method is used to further enhance the predictive ability of the model, and cross-validation is used to optimize hyperparameters, thereby improving the generalization ability of the model and helping to cope with complex and changeable clinical data.

[0020] The decision tree and support vector machine machine learning algorithms are used in combination with the boosting method to perform iterative weighted training and gradually correct the errors of the previous model. During the training process, cross-validation is used to optimize the model's hyperparameters. Specifically, the model formula is: f(X) represents the model's predicted output for the input feature vector X, t represents the number of steps in the current iteration, and α t represents the weight coefficient of the t-th weak learner, h t (X) represents the predicted output of the t-th weak learner for the input data X, T represents the number of iterations in the Boosting algorithm, P(R i |X) represents the i-th risk. When the input feature vector X is given, the predicted probability of the model output is f i (X) represents the prediction function of risk prediction model i, X represents the input feature vector, and i represents the index of risk type; During the model training process, the iterative weighted method is combined to gradually correct the model error, making the prediction model more accurate in predicting patient risks, especially in clinical environments with large amounts of data, effectively reducing the risk of overfitting and improving the stability and reliability of the prediction results.

[0021] Summarize the prediction results of the complication, comorbidity, and mortality risk models, combine the predicted probabilities of different risks, and form a comprehensive risk assessment report, which specifically includes: Based on Min-Max normalization, each probability value is normalized to between 0 and 1; The predicted probabilities of different risks are combined using a weighted average method to obtain a comprehensive risk score, assigning a weight to each risk; Based on the comprehensive risk score, thresholds are set to determine the patient's risk level, which is divided into low, medium and high levels. Based on each patient's comprehensive risk score and classification, a comprehensive risk assessment report is generated. The report content includes the patient's basic information, the individual prediction results of each risk model, the comprehensive risk score and recommended measures for risk assessment.

[0022] The predicted probabilities of different risks are combined using a weighted average method to obtain a comprehensive risk score. The weights assigned to each risk include: The comprehensive risk score calculation formula is: R T =l1·P(R1|X)+l2·P(R2|X)+l3·P(R3|X)+l i ·P(R i |X) Where R Tis the comprehensive risk score, P(R1|X) is the predicted probability of complication risk, l1 is the precursor of the predicted probability of complication risk, P(R2|X) is the predicted probability of complication risk, l2 is the weight of the predicted probability of complication risk, P(R3|X) is the predicted probability of death risk, l3 is the weight of the predicted probability of death risk, P(R i |X) is the probability of the i-th risk prediction, l i is the probability weight of the i-th risk prediction; By weighting each risk according to its characteristics and automatically adjusting the weights based on clinical experience, the contribution of different risk types to the overall score can be more reasonable. This innovative approach can effectively combine multiple risk assessments and enhance the predictive ability and decision support function of the overall score.

[0023] Based on the risk assessment results, a personalized early warning report is generated for each patient and pushed to medical staff in real time. It is also pushed to the doctor's workstation via a software plug-in in the hospital information system. Based on the severity of the risk, high-risk patients are given priority, assisting doctors in early intervention. Specific details include: Generate a personalized early warning report based on the assessment results, which includes the patient's risk type, the predicted probability of each risk and its risk level, recommended intervention measures, and priority classification; Through the real-time data push function within the hospital information system, the generated early warning report will be sent to relevant medical personnel in real time and pushed to the doctor's workstation through the software plug-in; Based on early warning reports, the system can provide decision-making support, automatically recommend intervention measures for high-risk patients, and provide best practice guidelines for early intervention; The system continuously tracks the patient's health status and updates risk assessments in real time; Real-time push of personalized early warning reports can not only reduce the intervention time of medical staff, but also improve the effectiveness of early intervention through system optimization suggestions. Based on risk priority classification, the system can intelligently push priority treatment suggestions for high-risk patients, further optimizing the allocation of medical resources.

[0024] Reference Figure 2 As shown in the figure, the patient safety management information system based on the medical quality control model includes: Data collection module, used to automatically capture patient medical history and unsafe incident data from the hospital information system; The risk assessment module uses a dual machine learning model to process and analyze the collected data to generate a prediction of the patient's risk of complications, comorbidities, and mortality; The result integration and report generation module is used to summarize the prediction results of each risk assessment model and generate a comprehensive risk assessment report; the early warning push module is used to push early warning information to the medical staff's workstation in real time for early intervention.

[0025] The risk assessment module specifically includes: Data pre-processing unit: Extracts patient medical records from the hospital information system, including historical conditions, physical examination data, treatment records, medication information, and unsafe event data. It then cleans, denoises, and standardizes the raw data to ensure data quality and consistency. Feature extraction unit: extracts key features and constructs these features into the format of model input; Data partitioning unit: divides the preprocessed data into training set, validation set and test set; Model training unit: uses decision tree and support vector machine machine learning algorithms, combined with boosting methods for model training, gradual weighted training, corrects the errors of the previous model, and optimizes the model's hyperparameters through cross-validation; Risk prediction unit: Based on the trained model, it predicts the medical records of new patients, outputs the prediction results of complications, comorbidities and death risks, makes independent predictions for different types of risks, and generates probability values ​​for each risk.

[0026] The result integration and report generation module specifically includes: Risk Assessment Summary Unit: Summarizes the prediction results of complications, comorbidities and mortality risk models into a comprehensive risk assessment report; Data normalization unit: Use the Min-Max normalization method to normalize the probability value of each risk prediction to between 0 and 1; Weighted average calculation unit: performs weighted average calculation on the predicted probabilities of different risks, assigns weights to each risk according to its importance, and obtains a comprehensive risk score; Risk level classification unit: Based on the comprehensive risk score, set thresholds to determine the patient's risk level and classify it; Report generation unit: Combines the patient's basic information, the prediction results of each risk model, the comprehensive risk score and the recommended intervention measures to generate a complete comprehensive risk assessment report.

[0027] Units included in the warning push module Real-time data push unit: Based on the patient's risk assessment results, personalized early warning reports are pushed to medical staff in real time. The push function within the hospital information system is used to transmit the reports directly to the doctor's workstation; Risk Priority Classification Unit: Prioritizes patients based on their comprehensive risk scores to ensure that high-risk patients receive priority treatment; Decision-making support unit: Based on risk assessment reports, it provides doctors with decision-making support, automatically recommends intervention measures, and helps doctors conduct early intervention; Health status monitoring unit: The system continuously tracks the patient's health status and updates risk assessments in real time.

[0028] In summary, the advantages of the present invention are: Improved risk prediction accuracy: This invention uses a data training method that combines dual machine learning algorithms with a boosting method to better process and analyze large amounts of medical record data, establish a high-precision risk prediction model, significantly improve the accuracy of predicting complications, comorbidities, and mortality risks, and reduce the occurrence of medical accidents. This method can generate personalized early warning reports in real time and push them to doctors' workstations through the hospital information system, enabling real-time intervention. This early warning can help doctors promptly identify high-risk patients and take appropriate intervention measures based on the predicted results, thereby reducing the probability of patients experiencing serious adverse events. Through multi-dimensional analysis of patient medical records and incorporating various risk prediction models, we generate comprehensive risk assessment reports for each patient and provide personalized intervention recommendations. This personalized risk assessment can help medical staff better understand the patient's health status and provide more precise medical services. By prioritizing patients based on comprehensive risk scores, medical resources can be allocated more efficiently, ensuring that high-risk patients receive priority attention and treatment. This optimized resource allocation not only improves medical efficiency but also enhances patient treatment outcomes. By continuously tracking patients' health status and providing real-time updated risk assessments, this invention provides medical personnel with a comprehensive decision-making support tool, helping doctors promptly identify potential risk points and avoid medical omissions and misdiagnoses. Combined with the automatic recommendation function of the machine learning model, this method not only provides doctors with real-time decision support, but also helps them develop more targeted treatment plans, thereby improving treatment effectiveness and patient satisfaction. This invention uses medical record data and unsafe event data in the hospital information system to perform big data analysis, providing hospital management with detailed risk assessment reports, and assisting hospitals in making data-driven decisions in overall medical safety management.

[0029] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions merely illustrate the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A patient safety management information method based on a medical quality control model, characterized by: include: Obtain patient condition data based on hospital information systems; The collected data is trained based on a dual machine learning algorithm combined with a boosting method to establish a high-precision risk prediction model. Based on the calculation of the risk prediction model, the patient's complication, comorbidity and mortality risk predictions are output; Summarize the model prediction results to form a comprehensive risk assessment report; Push early warning information to medical personnel in real time to assist in early intervention.

2. The patient safety management informationization method based on the medical quality control model according to claim 1 is characterized in that: The dual machine learning algorithm combined with the boosting method is used to train the collected data to establish a high-precision risk prediction model. Based on the calculation of the risk prediction model, the patient's complication, comorbidity risk and mortality risk prediction are output specifically including: Extract patient medical records from the hospital information system, including historical medical conditions, physical examination data, treatment records, medication information, and unsafe incident data, and clean, denoise, and standardize the raw data; Extract key features from medical record data, including patient age, gender, past medical history, current condition, and medication use, and construct these features into a model input format; Divide the preprocessed data into training set, validation set and test set; It uses two machine learning algorithms, decision tree and support vector machine, combined with boosting method, to iterate weighted training step by step, gradually correct the errors of the previous model. During the training process, cross-validation is used to optimize the model's hyperparameters. Apply the trained model to the medical records of new patients and calculate the risk of complications, comorbidities, and death based on the feature input; Make independent predictions for different types of risks and generate probability values ​​for each risk; Summarize the prediction results of complication, comorbidity and mortality risk models, combine the predicted probabilities of different risks, and form a comprehensive risk assessment report; Based on the risk assessment results, a personalized early warning report is generated for the patient, pushed to medical staff in real time, and pushed to the doctor's workstation through a software plug-in in the hospital information system. Based on the severity of the risk, high-risk patients are given priority to assist doctors in early intervention.

3. The patient safety management informationization method based on the medical quality control model according to claim 2 is characterized in that: The two machine learning algorithms, decision tree and support vector machine, are used in combination with boosting method to iterate weighted training step by step and gradually correct the error of the previous model. During the training process, cross-validation is used to optimize the hyperparameters of the model. include: The model formula is: f(X) represents the model's predicted output for the input feature vector X, t represents the number of steps in the current iteration, and α t represents the weight coefficient of the t-th weak learner, h t (X) represents the predicted output of the t-th weak learner for the input data X, T represents the number of iterations in the Boosting algorithm, P(R i |X) represents the i-th risk. When the input feature vector X is given, the predicted probability of the model output is f i (X) represents the prediction function of risk prediction model i, X represents the input feature vector, and i represents the index of risk type.

4. The patient safety management informationization method based on the medical quality control model according to claim 3 is characterized in that: The prediction results of the risk model for complications, comorbidities, and mortality are summarized, and the predicted probabilities of different risks are combined to form a comprehensive risk assessment report, which specifically includes: Based on Min-Max normalization, each probability value is normalized to between 0 and 1; The predicted probabilities of different risks are combined using a weighted average method to obtain a comprehensive risk score, assigning a weight to each risk; Based on the comprehensive risk score, thresholds are set to determine the patient's risk level, which is divided into low, medium and high levels. Based on each patient's comprehensive risk score and classification, a comprehensive risk assessment report is generated. The report content includes the patient's basic information, the individual prediction results of each risk model, the comprehensive risk score, and recommended measures for risk assessment.

5. The patient safety management informationization method based on the medical quality control model according to claim 4 is characterized in that: The weighted average method is used to combine the predicted probabilities of different risks to obtain a comprehensive risk score, and a specific weight is assigned to each risk. include: The comprehensive risk score calculation formula is: R T =l1·P(R1|X)+l2·P(R2|X)+l3·P(R3|X)+l i ·P(R i |X) Where R T is the comprehensive risk score, P(R1|X) is the predicted probability of complication risk, l1 is the precursor of the predicted probability of complication risk, P(R2|X) is the predicted probability of complication risk, l2 is the weight of the predicted probability of complication risk, P(R3|X) is the predicted probability of death risk, l3 is the weight of the predicted probability of death risk, P(R i |X) is the probability of the i-th risk prediction, l i is the probability weight of the i-th risk prediction.

6. The patient safety management informationization method based on the medical quality control model according to claim 5 is characterized in that: Based on the risk assessment results, a personalized early warning report is generated for each patient and pushed to medical staff in real time. The report is also pushed to the doctor's workstation via a software plug-in in the hospital information system. Based on the severity of the risk, high-risk patients are given priority, and doctors are assisted in early intervention. Specifically, the following are included: Generate a personalized early warning report based on the assessment results, which includes the patient's risk type, the predicted probability of each risk and its risk level, recommended intervention measures, and priority classification; Through the real-time data push function within the hospital information system, the generated early warning report will be sent to relevant medical personnel in real time and pushed to the doctor's workstation through the software plug-in; Based on early warning reports, the system can provide decision-making support, automatically recommend intervention measures for high-risk patients, and provide best practice guidelines for early intervention; The system continuously tracks the patient's health status and updates risk assessments in real time.

7. Patient safety management information system based on medical quality control model, characterized by: include: Data collection module, used to automatically capture patient medical history and unsafe incident data from the hospital information system; The risk assessment module uses a dual machine learning model to process and analyze the collected data to generate a prediction of the patient's risk of complications, comorbidities, and mortality; Result integration and report generation module, which is used to summarize the prediction results of each risk assessment model and generate a comprehensive risk assessment report; The early warning push module is used to push early warning information to the medical staff's workstation in real time for early intervention.

8. The patient safety management information system based on the medical quality control model according to claim 7 is characterized in that: The risk assessment module specifically includes: Data pre-processing unit: Extracts patient medical records from the hospital information system, including historical conditions, physical examination data, treatment records, medication information, and unsafe event data. It then cleans, denoises, and standardizes the raw data to ensure data quality and consistency. Feature extraction unit: extracts key features and constructs these features into the format of model input; Data partitioning unit: divides the preprocessed data into training set, validation set and test set; Model training unit: uses decision tree and support vector machine machine learning algorithms, combined with boosting methods for model training, gradual weighted training, corrects the errors of the previous model, and optimizes the model's hyperparameters through cross-validation; Risk prediction unit: Based on the trained model, it predicts the medical records of new patients, outputs the prediction results of complications, comorbidities and death risks, makes independent predictions for different types of risks, and generates probability values ​​for each risk.

9. The patient safety management information system based on the medical quality control model according to claim 8, characterized in that: The result integration and report generation module specifically includes: Risk Assessment Summary Unit: Summarizes the prediction results of complications, comorbidities and mortality risk models into a comprehensive risk assessment report; Data normalization unit: Use the Min-Max normalization method to normalize the probability value of each risk prediction to between 0 and 1; Weighted average calculation unit: performs weighted average calculation on the predicted probabilities of different risks, assigns weights to each risk according to its importance, and obtains a comprehensive risk score; Risk classification unit: Based on the comprehensive risk score, thresholds are set to determine the patient's risk level and classify them; Report generation unit: Generates a complete comprehensive risk assessment report by combining the patient's basic information, the prediction results of each risk model, the comprehensive risk score and the recommended intervention measures.

10. The patient safety management information system based on the medical quality control model according to claim 9, characterized in that: The warning push module includes the following units: Real-time data push unit: Based on the patient's risk assessment results, personalized early warning reports are pushed to medical staff in real time. The push function within the hospital information system is used to transmit the reports directly to the doctor's workstation; Risk Priority Classification Unit: Prioritizes patients based on their comprehensive risk scores to ensure that high-risk patients receive priority treatment; Decision-making support unit: Based on risk assessment reports, it provides doctors with decision-making support, automatically recommends intervention measures, and helps doctors conduct early intervention; Health status monitoring unit: The system continuously tracks the patient's health status and updates risk assessments in real time.

Images