CN119673452A - An acute kidney injury risk warning system - Google Patents
An acute kidney injury risk warning system Download PDFInfo
- Publication number
- CN119673452A CN119673452A CN202411737885.3A CN202411737885A CN119673452A CN 119673452 A CN119673452 A CN 119673452A CN 202411737885 A CN202411737885 A CN 202411737885A CN 119673452 A CN119673452 A CN 119673452A
- Authority
- CN
- China
- Prior art keywords
- patient
- data
- risk
- kidney
- early warning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses an acute kidney injury risk early warning system, and relates to the technical field of clinical decision support systems. The real-time monitoring and accurate prediction of the AKI risk of the patient are realized through data preprocessing, model construction and risk analysis, the early recognition rate is improved, a warning decision unit generates specific warning information according to the risk score and the kidney function change rule, provides targeted treatment suggestion, optimizes a warning strategy through a warning threshold adjustment and path matching module, improves the response speed of medical staff and the compliance of the patient, and a user interaction unit continuously optimizes a warning system through a feedback mechanism.
Description
Technical Field
The invention discloses an acute kidney injury risk early warning system, and relates to the technical field of clinical decision support systems.
Background
Currently, early identification and early warning of Acute Kidney Injury (AKI) is primarily dependent on the observation and analysis of clinical indicators by medical personnel. For example, the patent application with publication number CN118299054A discloses a critical patient acute kidney injury occurrence early warning system based on machine learning, which comprises a data acquisition and integration module, a feature screening and determination module, an optimal prediction model screening and determination module, a risk prediction result output module and a model interpretability analysis module, wherein the data acquisition and integration module acquires and preprocesses baseline features of patients entering an ICU, the feature screening and determination module is embedded with four sub-modules of baseline feature analysis, single factor logistic regression analysis, pelson/Szechwan correlation analysis and feature importance ranking, the optimal prediction model screening and determination module comprises an optimal variable number determination module and an optimal model determination module, the optimal prediction model screening and determination module comprises 2 sub-modules, and is used for determining the minimum variable number required by each machine learning model to achieve optimal prediction efficiency, and the model interpretability analysis module is used for interpreting the risk of AKI of each patient and contribution degree of key features based on an SHAP frame.
Although the risk of patient AKI can be estimated earlier and accurately, the method has certain limitation, needs a large amount of data preprocessing and model training, needs long-time data processing and model operation, causes inaccurate early warning time and lower real-time condition, cannot perform personalized early warning prompt for patient groups, and is easy to cause early warning fatigue and early warning desensitization.
Disclosure of Invention
The invention aims to provide an acute kidney injury risk early warning system, which remarkably improves the early recognition and intervention efficiency of AKI, reduces complications and mortality, improves clinical prognosis, optimizes medical resource allocation, promotes continuous improvement of medical quality, provides safer and more efficient medical services for patients, and solves the problems in the background technology.
In order to achieve the above object, the present invention provides a technical scheme, an acute kidney injury risk early warning system, comprising:
The data acquisition unit is used for collecting basic information, medical history, medication condition, laboratory examination result and biomarker data of a patient, distributing a unique identifier for the patient based on the basic information of the patient, and constructing an electronic health file of the patient;
The data analysis unit is used for preprocessing the acquired patient electronic health record, constructing an AKI risk assessment model, extracting laboratory examination results and biomarker data in the patient electronic health record, and predicting the risk of patient AKI occurrence based on the AKI risk assessment model;
the early warning decision unit is used for generating corresponding AKI early warning information according to the risk prediction result, adjusting an early warning threshold value by combining with the patient population characteristics, and matching the corresponding AKI early warning information with the notification path based on the early warning audience preference;
And the user interaction unit is used for displaying the AKI early warning information and the historical early warning record output by the early warning decision unit in real time, establishing a feedback mechanism and carrying out optimization and adjustment operation on the AKI early warning information based on the feedback mechanism.
Further, the data acquisition unit includes:
The multi-source data integration module is used for establishing multi-source heterogeneous data interfaces of each electronic medical record system, laboratory information system and medical imaging system of the hospital, extracting key information of a patient based on the multi-source heterogeneous data interfaces and unifying and standardizing data formats of the key information in different data sources;
the patient data packet generation module is used for classifying and packaging the acquired key information based on the data sources to generate patient data packets, and distributing a unique data packet identifier corresponding to the unique identification of the patient to each patient data packet;
And the electronic health record management module is used for carrying out association relation matching on basic data, medical history, medication condition, laboratory examination result and biomarker data of a patient based on the data source, and can simultaneously call corresponding patient data packets to construct the electronic health record of the patient when a unique identifier is input based on the association relation.
Further, the data analysis unit includes:
the data preprocessing module is used for carrying out data cleaning, normalization and standardization processing on the acquired electronic health record data of the patient, and extracting key features related to AKI risk assessment from the processed electronic health record data of the patient;
the model construction module is used for inputting the extracted key characteristics related to AKI risk assessment into a preset network model to obtain an AKI risk assessment model of each patient;
The risk analysis module is used for inputting the newly acquired electronic health record data of the patient into the AKI risk assessment model for risk identification and outputting a risk prediction result of the patient.
Further, the data analysis unit further includes:
Extracting laboratory examination results and biomarker data of a patient from the patient electronic health record, and extracting time sequence data of each data carried by the laboratory examination results and the biomarker data;
determining the renal function characteristics of the patient based on the time sequence data of each data, and determining the renal function change rule of the patient based on the time sequence data to obtain the change condition of the renal function characteristics of the patient in the electronic health record;
comparing the kidney function change rules of each patient, and integrating the patients with the similarity larger than or equal to a preset threshold value based on the comparison result to generate a similar patient group;
and identifying the renal function change rule of any patient in each patient group as a representative risk characteristic of the patient group as a patient group characteristic.
Further, the risk analysis module is further configured to:
the risk prediction result of the patient and the corresponding patient group characteristics are combined, the possibility of AKI of the patient is estimated, and quantification is carried out based on the estimation result, so that the risk score of the patient is obtained;
Comparing the risk score of the patient with a preset risk threshold, identifying a high risk patient based on the comparison result, and marking the high risk patient with high risk;
the historical risk recognition result of the patient is obtained, the long-term trend of the kidney function index of the patient is analyzed, and the long-term trend recognition based on the kidney function index can lead to early-stage signals of AKI.
Further, the kidney function change rule comprises a linear change rule that kidney function indexes are in a linear ascending or descending trend and represent the development of chronic kidney disease or recovery after treatment, a nonlinear change rule that kidney function indexes are irregular and represent acute kidney injury or repeated inflammatory reaction, a stepwise change rule that kidney function indexes are stable in certain time periods and are obviously changed in other time periods and can be related to treatment or life style change, and a threshold triggering rule that when the kidney function indexes exceed a specific threshold value, the kidney function is possibly obviously changed.
Further, the early warning decision unit includes:
the early warning information generation module is used for receiving the patient risk score and the kidney function change rule output by the risk analysis module, judging the early warning level according to the risk score and matching the corresponding AKI early warning information;
The early warning threshold adjustment module is used for extracting representative risk characteristics from each patient group, analyzing risk distribution conditions in each group, and adjusting an early warning threshold for each patient group according to the group characteristics and the risk distribution;
And the early warning path matching module is used for collecting AKI early warning information receiving preferences of medical staff and patients, customizing the content and format of the AKI early warning information according to the preferences and matching corresponding notification paths according to the early warning level.
Further, the early warning information generating module is further configured to:
Analyzing key factors possibly causing AKI, determining specific reasons of AKI, and matching corresponding laboratory examination and imaging examination items according to the specific reasons;
Providing corresponding treatment suggestions based on laboratory examination and imaging examination items, medical history of a patient and medication conditions of the patient, and generating AKI early warning information;
Medical resources are distributed according to AKI early warning information of a patient, treatment advice is analyzed through feedback data acquired by a user interaction unit, potential problems in a medical procedure are identified, and the treatment advice is adjusted.
Further, determining a renal function feature of the patient based on the time series data of each data, and obtaining a change condition of the renal function feature of the patient in the electronic health record based on the time series data, determining a renal function change rule of the patient, comprising:
determining multimodal development-work data for the patient's kidney over a plurality of time periods based on the time series data for each data;
Determining a working label of the patient in each period based on a preset multi-mode computing frame according to the multi-mode development-working data, and determining the expression level of the renal function of the patient according to the working label;
determining the renal function characteristics of the patient according to the expression level of the renal function of the patient, and acquiring the change condition of the renal function characteristics of the patient in the electronic health record based on the time sequence data;
defining urination frequency coefficient as design variable, taking kidney health body state as objective function, and motion index as constraint condition, and establishing evaluation model based on kidney function gradient;
Determining a time sequence data difference according to the change condition of the kidney function characteristics in the electronic health record, and determining urination related parameters and physical state related parameters of the patient according to the time sequence data;
Substituting the urination related parameter and the physical state related parameter into a renal function gradient assessment model to determine a renal function gradient change parameter of the patient;
Determining a kidney function gradient qualitative boundary range of the patient according to the gradient change parameters, and determining a kidney function level descending trend and a kidney function level leveling trend of the patient according to the gradient qualitative boundary range and parameters in the range;
Determining a quantitative change interval of the renal function of the patient according to the renal function level decline trend and the renal function level leveling trend through a preset logic quantitative rule;
And determining a threshold decreasing trend according to the quantitative change interval quantitative threshold condition, and determining the renal function change rule of the patient according to the threshold decreasing trend.
Further, after constructing the patient electronic health record, the method further comprises:
collecting a plurality of static gray-scale kidney images of a patient, screening out a target kidney image with the best quality, and importing the target kidney image into preset 3D graphic software to extract kidney form features;
Determining a selection operator according to the kidney form features, and determining healthy non-zero coefficients of a plurality of feature points of the kidney form features through the selection operator;
Calculating the kidney function injury degree of each patient according to healthy non-zero coefficients of a plurality of characteristic points:
Wherein A i is expressed as the kidney function injury degree of the ith patient, D i is expressed as the mental state index of the ith patient, ni is expressed as the number of the collection characteristic points of the ith patient, j is expressed as the j characteristic points, P j is expressed as the healthy non-zero coefficient of the j characteristic points, beta is expressed as a test factor, e is expressed as a natural constant, the value is 2.72, and F i is expressed as the historical kidney expression index of the ith patient;
Determining a renal function signature for each patient based on the extent of the injury to the renal function of the patient, and correlating the renal function signature with the electronic health record for the patient, the renal function signature comprising renal impairment, mild injury to the kidney, general injury to the kidney, and severe injury to the kidney;
Predictive assessment of risk of patient AKI occurrence is performed based on the renal function signature of each patient.
The invention discloses an acute kidney injury risk early warning system, which realizes real-time monitoring and accurate prediction of patient AKI risk through data preprocessing, model construction and risk analysis, improves early recognition rate, generates specific early warning information according to risk score and kidney function change rules, provides targeted treatment advice, optimizes early warning strategies through early warning threshold adjustment and approach matching modules, improves response speed of medical staff and compliance of patients, and continuously optimizes the early warning system through a feedback mechanism by a user interaction unit.
The technical scheme of the invention has the beneficial effects that:
The data analysis unit realizes real-time monitoring and accurate prediction of patient AKI risk through data preprocessing, model construction and risk analysis, improves early recognition rate, further realizes accurate recognition of patient kidney function characteristics and change rules through time sequence data analysis, provides an important data basis for AKI risk assessment and early warning, is beneficial to improving the early recognition rate and intervention efficiency of AKI, and the early warning decision unit generates specific early warning information according to risk scores and kidney function change rules and provides targeted treatment advice, and simultaneously optimizes an early warning strategy through an early warning threshold adjustment and path matching module, reduces false alarm and omission, improves response speed of medical staff and patient compliance, and the user interaction unit realizes real-time display and historical record analysis of early warning information and continuously optimizes the early warning system through a feedback mechanism.
The kidney function damage degree of each patient can be calculated according to the daily mental state and the kidney performance condition of each patient, the kidney health condition of each patient can be rapidly, objectively and accurately determined, and further unhealthy kidney patients can be rapidly screened out to be subjected to the prior AKI evaluation, so that the working efficiency and the accuracy of judging the kidney function damage are improved.
With the aging of population and the increase of chronic patients, the incidence of AKI is on the rising trend, and early identification and intervention have important significance for reducing mortality and improving prognosis of patients. The invention can realize real-time monitoring and accurate prediction, can effectively improve the early recognition rate and the treatment success rate of AKI, solves the problems of inaccurate early warning time, low real-time performance, lack of personalized early warning and the like in the prior art, meets clinical requirements, and has wide market prospect.
The invention can early identify and intervene AKI, reduce the consumption of medical resources such as Intensive Care Unit (ICU) bed use, dialysis and the like, and automatically early warning and individuation treatment advice, effectively reduce the treatment cost and the hospitalization time of AKI patients, reduce the medical cost, improve the working efficiency of medical staff and reduce complications and mortality, thereby generating remarkable economic benefit.
Drawings
Fig. 1is a diagram of an acute kidney injury risk early warning system according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order to solve the technical problems that in the prior art, a large amount of data preprocessing and model training are required, a long time of data processing and model operation are required, the situation that early warning time is inaccurate and real-time is low is caused, and personalized early warning prompt cannot be carried out for a patient group, early warning fatigue and early warning desensitization are easy to cause, referring to fig. 1, the embodiment provides the following technical scheme:
an acute kidney injury risk early warning system, comprising:
The data acquisition unit is used for collecting basic information (such as name, identity card number and the like), medical history, medication condition, laboratory examination result (such as creatinine, urine volume and the like) and biomarker data (such as NGAL, KIM-1 and the like) of a patient, distributing a unique identifier for the patient based on the basic information of the patient, constructing an electronic health file of the patient, and recording the medical history, medication condition, laboratory examination result and biomarker data of the patient into the electronic health file of the patient;
The data analysis unit is used for preprocessing the acquired patient electronic health record, constructing an AKI risk assessment model, extracting laboratory examination results and biomarker data in the patient electronic health record, predicting the risk of patient AKI based on the AKI risk assessment model, monitoring the dynamic change of the renal function of the patient in real time, automatically identifying the descending trend of the renal function, and triggering an early warning mechanism;
The early warning decision unit is used for generating corresponding AKI early warning information according to the risk prediction result, including early warning level, possible reasons, recommended examination items and preliminary treatment suggestions, adjusting an early warning threshold value by combining with patient group characteristics, matching the corresponding AKI early warning information and notification paths based on early warning audience preference, optimizing an early warning strategy, and reducing false alarm and missing report;
And the user interaction unit is used for displaying the AKI early warning information and the historical early warning record which are output by the early warning decision unit in real time, including past early warning events, treatment measures and patient fatalities, establishing a feedback mechanism, carrying out optimization and adjustment operation on the AKI early warning information based on the feedback mechanism, evaluating the effect of different intervention measures on reducing AKI risk according to the feedback mechanism, comparing risk scores before and after intervention to measure treatment effect, monitoring the change of AKI incidence rate, and evaluating the effectiveness of a prevention strategy.
In the embodiment, through constructing the patient electronic health record, not only is a rich data basis provided for the evaluation of AKI risk, but also the integrity and traceability of patient information are ensured, and the doctor is facilitated to make more accurate decisions in the diagnosis and treatment process, and an efficient and accurate AKI risk evaluation model is constructed, so that the risk of patient AKI is accurately predicted, the dynamic change of renal function is monitored in real time, the early recognition rate is improved, the early recognition and intervention efficiency of AKI can be remarkably improved, the complications and mortality of patients caused by AKI are reduced, the clinical prognosis is improved, in addition, the early warning threshold and notification path can be adjusted according to the patient population characteristics and the early warning audience preference, the false report and missing report are reduced, and the personalized early warning mechanism can ensure that medical staff can obtain key information in the shortest time, thereby taking timely and effective intervention measures are facilitated to promote the development and application of a clinical decision support system, and the overall quality and level of medical services are improved.
In this embodiment, the data acquisition unit includes:
the multi-source data integration module is used for establishing multi-source heterogeneous data interfaces of each electronic medical record system, laboratory information system and medical imaging system of a hospital, extracting key information of a patient based on the multi-source heterogeneous data interfaces, including basic information, medical history, medication condition, laboratory examination results and biomarker data, and unifying and standardizing data formats of the key information in different data sources;
The system comprises a patient data packet generation module, a data packet generation module and a data processing module, wherein the patient data packet generation module is used for classifying and packaging acquired key information based on data sources to generate patient data packets, and each data packet comprises related information of a patient in different systems;
And the electronic health record management module is used for carrying out association relation matching on basic data, medical history, medication condition, laboratory examination result and biomarker data of a patient based on the data source, and can simultaneously call corresponding patient data packets to construct the electronic health record of the patient when a unique identifier is input based on the association relation.
In this embodiment, by uniformly accessing and processing each system of the hospital, the key information of the patient is more effectively extracted from the multi-source heterogeneous data, the key information such as the disease name, the symptom description, the treatment scheme and the like is extracted from the unstructured or semi-structured text in the EMR, the required data items such as the blood creatinine value, the urine volume data, the image diagnosis result and the like are directly extracted aiming at the structured data in the LIS and the PACS, the association relationship between the data is established, the medical history in the EMR and the laboratory examination result in the LIS are associated through the patient ID, the perfect patient electronic health record is constructed, and a solid data base is provided for subsequent AKI risk assessment and early warning.
In this embodiment, the data analysis unit includes:
The data preprocessing module is used for carrying out data cleaning, normalization and standardization processing on the acquired electronic health record data of the patient, and extracting key characteristics related to AKI risk assessment from the processed electronic health record data of the patient, wherein the key characteristics comprise time sequence characteristics, renal function characteristics and the like;
the model construction module is used for inputting the extracted key characteristics related to AKI risk assessment into a preset network model to obtain an AKI risk assessment model of each patient;
the risk analysis module is used for inputting the newly acquired electronic health record data of the patient into the AKI risk assessment model for risk identification, outputting risk prediction results of the patient, including risk level, risk probability and other forms of assessment results, and further used for:
the risk prediction result of the patient and the corresponding patient group characteristics are combined, the possibility of AKI of the patient is estimated, and quantification is carried out based on the estimation result, so that the risk score of the patient is obtained;
Comparing the risk score of the patient with a preset risk threshold, identifying a high risk patient based on the comparison result, and marking the high risk patient with high risk;
Acquiring a historical risk recognition result of the patient, analyzing a long-term trend of the kidney function index of the patient, and recognizing an early signal which possibly causes AKI based on the long-term trend of the kidney function index, such as a trend of gradually decreasing kidney function;
In the embodiment, through combining the risk prediction result and the patient group characteristics, the possibility of the patient to generate AKI is estimated, and quantified into a risk score, medical staff takes corresponding precautionary measures according to the score, so that the medical staff is helped to quickly identify the potential high-risk patient, and the high-risk patient is ensured to be preferentially concerned and intervened in time, thereby the occurrence of AKI is possibly reduced, the early detection rate of AKI is improved, the long-term trend of the kidney function index is monitored, the disease progress of the patient can be better understood, the basis is provided for personalized treatment, and the quality and efficiency of medical service are improved;
in this embodiment, the data analysis unit further includes:
Extracting laboratory examination results and biomarker data of a patient from the patient electronic health record, and extracting time sequence data of each data carried by the laboratory examination results and the biomarker data;
Determining the renal function characteristics of the patient based on the time sequence data of each data, including stable renal function, gradual decline of renal function, acute injury of renal function, fluctuation of renal function, and the like, and determining the renal function change rule of the patient based on the time sequence data by acquiring the change condition of the renal function characteristics of the patient in the electronic health record;
In this embodiment, the kidney function change rule includes a linear change rule that kidney function index is in a straight line ascending or descending trend, representing development of chronic kidney disease or recovery after treatment, a nonlinear change rule that kidney function index is changed irregularly, representing acute kidney injury or repeated inflammatory reaction, a stepwise change rule that kidney function index is stable in some time periods and changes significantly in other time periods, possibly related to treatment or life style change, a threshold trigger rule that indicates that kidney function may change significantly when kidney function index exceeds a specific threshold;
comparing the kidney function change rules of each patient, and integrating the patients with the similarity larger than or equal to a preset threshold value based on the comparison result to generate a similar patient group;
and identifying the renal function change rule of any patient in each patient group as a representative risk characteristic of the patient group as a patient group characteristic.
In the embodiment, key data such as kidney function indexes and vital signs of a patient are monitored in real time, an AKI risk prediction model is constructed to comprehensively analyze various indexes of the patient, intelligent and accurate prediction of AKI risk is achieved, early identification capacity of AKI by a medical institution is improved, powerful support is provided for clinical decision, morbidity and mortality of AKI are remarkably reduced, the patient group is deeply understood through analysis of kidney function change rules of the patient and classification of the patient group, change modes of kidney functions of the patient are understood, basis is provided for personalized treatment, treatment strategies are optimized, pertinence and efficiency of treatment are improved, and development of personalized treatment is promoted.
In this embodiment, the early warning decision unit includes:
the early warning information generation module is used for receiving the patient risk score and the kidney function change rule output by the risk analysis module, judging the early warning level according to the risk score, matching the corresponding AKI early warning information and further used for:
Analyzing key factors that may lead to AKI, such as drug use, surgery, infection, dehydration, etc., determining specific causes of AKI in order to take targeted therapeutic measures and matching corresponding laboratory and imaging exam items according to the specific causes;
providing corresponding treatment suggestions, such as medication adjustment, fluid infusion, dialysis and the like, based on laboratory examination and imaging examination items, medical history of a patient and medication conditions of the patient, and generating AKI early warning information;
Medical resources such as Intensive Care Unit (ICU) beds, dialysis resources and the like are distributed according to AKI early warning information of patients, treatment suggestions are analyzed through feedback data obtained by a user interaction unit, potential problems in a medical process are identified, the treatment suggestions are adjusted, and quality improvement measures are promoted;
The early warning threshold adjustment module is used for extracting representative risk characteristics from each patient group, analyzing risk distribution conditions in each group, and adjusting an early warning threshold for each patient group according to the group characteristics and the risk distribution, for example, for a high risk group, reducing the threshold to improve sensitivity, and for a low risk group, improving the threshold to reduce false alarm;
And the early warning path matching module is used for collecting AKI early warning information receiving preferences of medical staff and patients, customizing the content and format of the AKI early warning information according to the preferences and matching corresponding notification paths according to the early warning level.
In this embodiment, specific AKI early warning information is generated according to real-time data and risk scores of patients, key factors of AKI are analyzed and targeted treatment advice is provided, so that the patients are ensured to get timely and effective intervention, the problems of early warning fatigue and early warning desensitization are effectively solved by dynamically adjusting early warning threshold values and individuation matching notification paths, response speed of medical staff to early warning information and compliance of the patients are improved, early recognition rate and treatment success rate of AKI are remarkably improved, health risks of the patients are reduced, medical resource allocation is optimized, medical quality is promoted to be continuously improved, and safer and more efficient medical services are provided for the patients.
In one embodiment, determining a renal function feature of the patient based on the time series data for each data, and obtaining a change in the renal function feature of the patient in the electronic health record based on the time series data, determining a renal function change rule for the patient, comprises:
determining multimodal development-work data for the patient's kidney over a plurality of time periods based on the time series data for each data;
Determining a working label of the patient in each period based on a preset multi-mode computing frame according to the multi-mode development-working data, and determining the expression level of the renal function of the patient according to the working label;
determining the renal function characteristics of the patient according to the expression level of the renal function of the patient, and acquiring the change condition of the renal function characteristics of the patient in the electronic health record based on the time sequence data;
defining urination frequency coefficient as design variable, taking kidney health body state as objective function, and motion index as constraint condition, and establishing evaluation model based on kidney function gradient;
Determining a time sequence data difference according to the change condition of the kidney function characteristics in the electronic health record, and determining urination related parameters and physical state related parameters of the patient according to the time sequence data;
Substituting the urination related parameter and the physical state related parameter into a renal function gradient assessment model to determine a renal function gradient change parameter of the patient;
Determining a kidney function gradient qualitative boundary range of the patient according to the gradient change parameters, and determining a kidney function level descending trend and a kidney function level leveling trend of the patient according to the gradient qualitative boundary range and parameters in the range;
Determining a quantitative change interval of the renal function of the patient according to the renal function level decline trend and the renal function level leveling trend through a preset logic quantitative rule;
And determining a threshold decreasing trend according to the quantitative change interval quantitative threshold condition, and determining the renal function change rule of the patient according to the threshold decreasing trend.
The technical scheme has the beneficial effects that the working level of the kidney of the patient can be intuitively determined by determining the descending trend of the kidney function level and the leveling trend of the kidney function level of the patient, so that the quantitative threshold change interval of the kidney function of the patient can be rapidly determined according to the quantitative rule, and the kidney function change rule is determined according to the data change trend. And the judging efficiency and accuracy are improved.
In one embodiment, after constructing the patient electronic health record, further comprising:
collecting a plurality of static gray-scale kidney images of a patient, screening out a target kidney image with the best quality, and importing the target kidney image into preset 3D graphic software to extract kidney form features;
Determining a selection operator according to the kidney form features, and determining healthy non-zero coefficients of a plurality of feature points of the kidney form features through the selection operator;
Calculating the kidney function injury degree of each patient according to healthy non-zero coefficients of a plurality of characteristic points:
Wherein A i is expressed as the kidney function injury degree of the ith patient, D i is expressed as the mental state index of the ith patient, ni is expressed as the number of the collection characteristic points of the ith patient, j is expressed as the j characteristic points, P j is expressed as the healthy non-zero coefficient of the j characteristic points, beta is expressed as a test factor, e is expressed as a natural constant, the value is 2.72, and F i is expressed as the historical kidney expression index of the ith patient;
Determining a renal function signature for each patient based on the extent of the injury to the renal function of the patient, and correlating the renal function signature with the electronic health record for the patient, the renal function signature comprising renal impairment, mild injury to the kidney, general injury to the kidney, and severe injury to the kidney;
Predictive assessment of risk of patient AKI occurrence is performed based on the renal function signature of each patient.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should be covered by the protection scope of the present invention by making equivalents and modifications to the technical solution and the inventive concept thereof.
Claims (10)
1. An acute kidney injury risk early warning system, comprising:
The data acquisition unit is used for collecting basic information, medical history, medication condition, laboratory examination result and biomarker data of a patient, distributing a unique identifier for the patient based on the basic information of the patient, and constructing an electronic health file of the patient;
The data analysis unit is used for preprocessing the acquired patient electronic health record, constructing an AKI risk assessment model, extracting laboratory examination results and biomarker data in the patient electronic health record, and predicting the risk of patient AKI occurrence based on the AKI risk assessment model;
the early warning decision unit is used for generating corresponding AKI early warning information according to the risk prediction result, adjusting an early warning threshold value by combining with the patient population characteristics, and matching the corresponding AKI early warning information with the notification path based on the early warning audience preference;
And the user interaction unit is used for displaying the AKI early warning information and the historical early warning record output by the early warning decision unit in real time, establishing a feedback mechanism and carrying out optimization and adjustment operation on the AKI early warning information based on the feedback mechanism.
2. An acute kidney injury risk warning system according to claim 1, wherein the data acquisition unit comprises:
The multi-source data integration module is used for establishing multi-source heterogeneous data interfaces of each electronic medical record system, laboratory information system and medical imaging system of the hospital, extracting key information of a patient based on the multi-source heterogeneous data interfaces and unifying and standardizing data formats of the key information in different data sources;
the patient data packet generation module is used for classifying and packaging the acquired key information based on the data sources to generate patient data packets, and distributing a unique data packet identifier corresponding to the unique identification of the patient to each patient data packet;
And the electronic health record management module is used for carrying out association relation matching on basic data, medical history, medication condition, laboratory examination result and biomarker data of a patient based on the data source, and can simultaneously call corresponding patient data packets to construct the electronic health record of the patient when a unique identifier is input based on the association relation.
3. An acute kidney injury risk warning system according to claim 1, wherein the data analysis unit comprises:
the data preprocessing module is used for carrying out data cleaning, normalization and standardization processing on the acquired electronic health record data of the patient, and extracting key features related to AKI risk assessment from the processed electronic health record data of the patient;
the model construction module is used for inputting the extracted key characteristics related to AKI risk assessment into a preset network model to obtain an AKI risk assessment model of each patient;
The risk analysis module is used for inputting the newly acquired electronic health record data of the patient into the AKI risk assessment model for risk identification and outputting a risk prediction result of the patient.
4. The acute kidney injury risk warning system according to claim 1, wherein the data analysis unit further comprises:
Extracting laboratory examination results and biomarker data of a patient from the patient electronic health record, and extracting time sequence data of each data carried by the laboratory examination results and the biomarker data;
determining the renal function characteristics of the patient based on the time sequence data of each data, and determining the renal function change rule of the patient based on the time sequence data to obtain the change condition of the renal function characteristics of the patient in the electronic health record;
comparing the kidney function change rules of each patient, and integrating the patients with the similarity larger than or equal to a preset threshold value based on the comparison result to generate a similar patient group;
and identifying the renal function change rule of any patient in each patient group as a representative risk characteristic of the patient group as a patient group characteristic.
5. The acute kidney injury risk early warning system according to claim 1, wherein the risk analysis module is further configured to:
the risk prediction result of the patient and the corresponding patient group characteristics are combined, the possibility of AKI of the patient is estimated, and quantification is carried out based on the estimation result, so that the risk score of the patient is obtained;
Comparing the risk score of the patient with a preset risk threshold, identifying a high risk patient based on the comparison result, and marking the high risk patient with high risk;
the historical risk recognition result of the patient is obtained, the long-term trend of the kidney function index of the patient is analyzed, and the long-term trend recognition based on the kidney function index can lead to early-stage signals of AKI.
6. An acute kidney injury risk warning system according to claim 1, wherein the renal function change rule comprises:
the linear change rule is that the kidney function index is in a linear ascending or descending trend and represents the development of chronic kidney disease or recovery after treatment;
Nonlinear change rules, namely irregular change of kidney function index, representing acute kidney injury or repeated inflammatory reaction;
A rule of stepwise change that the renal function index is stable for some periods of time, while it is significantly changed for other periods of time, possibly related to treatment or lifestyle changes;
threshold triggering rules, when the kidney function indicator exceeds a certain threshold, indicate that a significant change in kidney function may occur.
7. An acute kidney injury risk early warning system according to claim 1, wherein the early warning decision unit comprises:
the early warning information generation module is used for receiving the patient risk score and the kidney function change rule output by the risk analysis module, judging the early warning level according to the risk score and matching the corresponding AKI early warning information;
The early warning threshold adjustment module is used for extracting representative risk characteristics from each patient group, analyzing risk distribution conditions in each group, and adjusting an early warning threshold for each patient group according to the group characteristics and the risk distribution;
And the early warning path matching module is used for collecting AKI early warning information receiving preferences of medical staff and patients, customizing the content and format of the AKI early warning information according to the preferences and matching corresponding notification paths according to the early warning level.
8. The acute kidney injury risk warning system according to claim 1, wherein the warning information generation module is further configured to:
Analyzing key factors possibly causing AKI, determining specific reasons of AKI, and matching corresponding laboratory examination and imaging examination items according to the specific reasons;
Providing corresponding treatment suggestions based on laboratory examination and imaging examination items, medical history of a patient and medication conditions of the patient, and generating AKI early warning information;
Medical resources are distributed according to AKI early warning information of a patient, treatment advice is analyzed through feedback data acquired by a user interaction unit, potential problems in a medical procedure are identified, and the treatment advice is adjusted.
9. An acute kidney injury risk warning system according to claim 1, wherein determining the patient's renal function characteristics based on the time series data for each data, and obtaining the patient's renal function characteristics change in the electronic health record based on the time series data, determining the patient's renal function change rules, comprises:
determining multimodal development-work data for the patient's kidney over a plurality of time periods based on the time series data for each data;
Determining a working label of the patient in each period based on a preset multi-mode computing frame according to the multi-mode development-working data, and determining the expression level of the renal function of the patient according to the working label;
determining the renal function characteristics of the patient according to the expression level of the renal function of the patient, and acquiring the change condition of the renal function characteristics of the patient in the electronic health record based on the time sequence data;
defining urination frequency coefficient as design variable, taking kidney health body state as objective function, and motion index as constraint condition, and establishing evaluation model based on kidney function gradient;
Determining a time sequence data difference according to the change condition of the kidney function characteristics in the electronic health record, and determining urination related parameters and physical state related parameters of the patient according to the time sequence data;
Substituting the urination related parameter and the physical state related parameter into a renal function gradient assessment model to determine a renal function gradient change parameter of the patient;
Determining a kidney function gradient qualitative boundary range of the patient according to the gradient change parameters, and determining a kidney function level descending trend and a kidney function level leveling trend of the patient according to the gradient qualitative boundary range and parameters in the range;
Determining a quantitative change interval of the renal function of the patient according to the renal function level decline trend and the renal function level leveling trend through a preset logic quantitative rule;
And determining a threshold decreasing trend according to the quantitative change interval quantitative threshold condition, and determining the renal function change rule of the patient according to the threshold decreasing trend.
10. The acute kidney injury risk warning system of claim 1, further comprising, after constructing the patient electronic health record:
collecting a plurality of static gray-scale kidney images of a patient, screening out a target kidney image with the best quality, and importing the target kidney image into preset 3D graphic software to extract kidney form features;
Determining a selection operator according to the kidney form features, and determining healthy non-zero coefficients of a plurality of feature points of the kidney form features through the selection operator;
Calculating the kidney function injury degree of each patient according to healthy non-zero coefficients of a plurality of characteristic points:
Wherein A i is expressed as the kidney function injury degree of the ith patient, D i is expressed as the mental state index of the ith patient, ni is expressed as the number of the collection characteristic points of the ith patient, j is expressed as the j characteristic points, P j is expressed as the healthy non-zero coefficient of the j characteristic points, beta is expressed as a test factor, e is expressed as a natural constant, the value is 2.72, and F i is expressed as the historical kidney expression index of the ith patient;
Determining a renal function signature for each patient based on the extent of the injury to the renal function of the patient, and correlating the renal function signature with the electronic health record for the patient, the renal function signature comprising renal impairment, mild injury to the kidney, general injury to the kidney, and severe injury to the kidney;
Predictive assessment of risk of patient AKI occurrence is performed based on the renal function signature of each patient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411737885.3A CN119673452A (en) | 2024-11-29 | 2024-11-29 | An acute kidney injury risk warning system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411737885.3A CN119673452A (en) | 2024-11-29 | 2024-11-29 | An acute kidney injury risk warning system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119673452A true CN119673452A (en) | 2025-03-21 |
Family
ID=94989999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411737885.3A Pending CN119673452A (en) | 2024-11-29 | 2024-11-29 | An acute kidney injury risk warning system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119673452A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120413079A (en) * | 2025-04-25 | 2025-08-01 | 北京大学人民医院 | A CRRT unplanned discontinuation warning and decision support information system |
| CN120748650A (en) * | 2025-08-29 | 2025-10-03 | 西安大兴医院 | Comprehensive nursing management system based on big data analysis |
-
2024
- 2024-11-29 CN CN202411737885.3A patent/CN119673452A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120413079A (en) * | 2025-04-25 | 2025-08-01 | 北京大学人民医院 | A CRRT unplanned discontinuation warning and decision support information system |
| CN120413079B (en) * | 2025-04-25 | 2025-10-03 | 北京大学人民医院 | CRRT (China radio remote control) unscheduled off-line early warning and decision support information system |
| CN120748650A (en) * | 2025-08-29 | 2025-10-03 | 西安大兴医院 | Comprehensive nursing management system based on big data analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN119673452A (en) | An acute kidney injury risk warning system | |
| CN118173253B (en) | System and method for analyzing and managing based on patient data | |
| KR20090024808A (en) | Assessment of dementia and dementia disorders | |
| JP2018524137A (en) | Method and system for assessing psychological state | |
| CN112786203A (en) | Machine learning diabetic retinopathy morbidity risk prediction method and application | |
| CN119227784B (en) | A decision tree generation method based on large language model and related equipment | |
| CN118280607B (en) | Intelligent critical index monitoring method and system based on clinical data | |
| CN112183572A (en) | Method and device for generating prediction model for predicting pneumonia severity | |
| CN117219275A (en) | Child drug-induced liver injury risk identification prediction method and system | |
| CN119541752A (en) | An intelligent recommendation system for ear, nose and throat care solutions based on cloud computing | |
| CN119028512A (en) | A real-time obstetric nursing optimization method based on big data | |
| Chen et al. | A deep neural network-based model for OSA severity classification using unsegmented peripheral oxygen saturation signals | |
| CN119153121A (en) | Orthopedics patient data management and remote monitoring platform | |
| CN119541875A (en) | A method and system for predicting sepsis deterioration based on TCM big data | |
| Panga | Advanced data mining techniques in healthcare: Bridging AI and big data | |
| CN118471420A (en) | Health physical examination report data analysis management system based on Internet | |
| CN117594227A (en) | Health status monitoring methods, devices, media and equipment based on wearable devices | |
| CN117653053A (en) | A way to predict health risks with smartwatches | |
| Seo et al. | A deep neural network based Wake-After-Sleep-Onset time aware sleep apnea severity Estimation scheme using Single-Lead ECG data | |
| CN117612731A (en) | Method for constructing chronic disease risk model | |
| CN119385526B (en) | Intelligent diagnosis wearable device for acute myocardial infarction based on structural causal model | |
| US20240266062A1 (en) | Disease risk evaluation method, disease risk evaluation system, and health information processing device | |
| CN119380969B (en) | A safe diagnostic method for medical monitoring | |
| CN119964841A (en) | A risk detection method and system for adverse events in drug clinical trials | |
| CN112185554A (en) | Method and system for analyzing and predicting health condition based on health information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |