39 results about "Patient classification" patented technology

The invention discloses an Alzheimer's disease rehabilitation assistance system based on artificial intelligence. The Alzheimer's disease rehabilitation assistance system comprises a data collection module, a cognitive evaluation module, a patient classification module, a rehabilitation planning module and a comprehensive assistance module. The invention belongs to the technical field of Alzheimer's disease rehabilitation management, and particularly relates to an Alzheimer's disease rehabilitation auxiliary system based on artificial intelligence, which realizes cognitive risk level judgment by acquiring multi-modal data of voice, writing, behavior and nursing evaluation, extracting structural features and constructing a cognitive evaluation method fusing rules and a lightweight model; performing multi-dimensional symptom grading in combination with patient behaviors and living abilities, screening personalized intervention tasks through an adaptive scoring mechanism, and generating a dynamic rehabilitation path and plan; closed-loop linkage among cognitive assessment, hierarchical classification and rehabilitation intervention is realized, and the accuracy and practicability of rehabilitation assistance are improved.

The invention relates to the technical field of drug purchase quantity prediction, in particular to an intelligent drug purchase quantity prediction method using medical insurance consumption data, comprising the following steps: S1, acquiring historical medical insurance consumption data including drug names, sales quantity, sales time, medical institution information and patient classification information; s2, preprocessing the medical insurance consumption data, including data cleaning, missing value filling and abnormal value correction, to form a standardized data set; and S3, based on the standardized data set, a drug demand feature matrix X is constructed, and each row represents a drug consumption record of one time period. According to the intelligent drug purchase quantity prediction method based on the medical insurance consumption data, trend, seasonal and random fluctuations are decomposed and separated through a time sequence, and noise interference is reduced; and combining an XGBoost and LSTM integrated model, comprehensively capturing a linear and nonlinear relationship, reducing a final prediction error rate, dynamically calculating a safe inventory threshold value, and avoiding excessive stockpiling of drugs.

Disclosed herein are classifier models, computer implemented systems, machine learning systems and methods thereof for classifying asymptomatic patients into a risk category for having or developing cancer and / or classifying a patient with an increased risk of having or developing cancer into an organ system-based malignancy class membership and / or into a specific cancer class membership.

Provided is a process including: accessing clinical trial data of an ongoing clinical trial that is not yet complete, the clinical trial having a plurality of treatment groups and a plurality of patients in the treatment groups; detecting, with an anomaly detection model, an anomaly in the clinical trial data for a first patient, the anomaly corresponding to a first patient among the plurality of patients; in response to detecting the anomaly classifying the first patient as anomalous in the trial; generating, with the computer system, by an artificial intelligence model, a candidate explanation for the detected anomaly by designating data in a record for the first patient as potentially correlated with the anomaly; and causing the anomaly and the candidate explanation to be presented to a data safety monitoring board (“DSMB”) of the clinical trial.

A system and method for predicting the Continuous Cardiac Output (CCO) of patients is disclosed. The method includes receiving a request from one or more patients and one or medical professionals and determining a set of recovery patterns of the one or more patients. The method further classifying the one or more patients into one or more predefined profiles and predicting the CCO of the one or more patients based on the request, the set of recovery patterns and result of classification using health management based AI model. Further, the method includes generating recommendations corresponding to the predicted CCO and outputting the predicted CCO of the one or more patients and the generated recommendations along with insights on user interface screen of electronic devices associated with the one or more patients and the one or more medical professionals, to take corrective actions in real time and future.

This invention relates to the field of medical information management technology, specifically to a mobile tent hospital medical information management system and method based on WeChat mini-programs. The system includes the following steps: a mini-program login verification module, a treatment process recording module, a patient classification and bed allocation module, a network topology management module, and a rapid medical record export module. In this invention, abnormal behavior is identified by analyzing user login behavior data, and the identity verification process is adjusted to raise the security threshold for data access and strengthen the defense against data leakage incidents. RFID tags are used to achieve automatic identification of patient information and real-time updates of electronic medical records, improving the response efficiency of temporary medical facilities and ensuring the continuity and efficiency of medical services in disaster response scenarios. Real-time monitoring and optimization of the network topology ensures smooth data transmission within the medical network. Combined with the analysis and formatting of electronic medical record text content, rapid export of electronic medical records is achieved.

The application discloses a patient classification method and device based on magnetic resonance parameters, equipment, medium, the method comprises the following steps: obtaining the target magnetic resonance parameter of the target patient; the target magnetic resonance parameter is selected from a plurality of preset magnetic resonance parameters based on the cardiac resynchronization label of the sample patient and the true result of the occurrence of the preset endpoint event; the target magnetic resonance parameter comprises a myocardial strain parameter, a gadolinium contrast agent delayed enhancement result and a left ventricular torsion parameter; the target classification tree is input into the target patient, and the target patient is classified by using the target classification tree, and the target classification result corresponding to the target patient is obtained; the target classification tree is obtained by extending the preset classification tree based on the cut-off value of each target magnetic resonance parameter and the influence of meeting each cut-off value on cardiac resynchronization; and the target classification result is output. The application can classify patients more accurately.

The vehicle first-aid medical method for guiding first-aid patient classification, treatment and hospital transfer according to the present application includes the following steps: a patient information registration unit receives necessary information about a patient in advance before getting on a vehicle or in the vehicle after getting on the vehicle; a patient symptom input unit receives the current symptoms of the patient in multiple ways; a first-aid treatment guideline providing unit provides a first-aid treatment method suitable for the patient based on the collected information; and a remote medical support unit supports real-time communication with a remote medical staff in the vehicle, so that the patient or a first-aid team member can obtain direct advice and guidance from a first-aid medical specialist.

Identifying the causes of haematuria requires detailed clinical investigations. Unbalanced patient populations (mostly male) and diverse causes complicate patient classification and biomarker identification. Multi-marker analysis with advanced methods can improve diagnosis even in unbalanced datasets. We used k-means clustering, decision trees, random forest with LIME explainer, and CACTUS algorithm to classify patients as healthy (with no clear cause of haematuria) or sick (with an identified cause of haematuria e.g., bladder cancer, or infection). The CACTUS algorithm was more robust to dataset imbalance, achieving balanced accuracy of 0.747 for both genders, 0.718 for females, and 0.803 for males. Key biomarkers identified included microalbumin, male gender, and tPSA for the whole dataset; age, microalbumin, tPSA, cystatin C, BTA, HAD, and S100A4 for males; and microalbumin, IL-8, pERK, and CXCL16 for females. CACTUS outperformed decision trees and random forest, highlighting potential novel biomarkers for future diagnosis.

The application provides an AI blood collection decompression grip ware system and an application method, and the system comprises a sensor matrix module, a main control processing module connected with the sensor matrix module, a voice interaction module connected with the main control processing module, and a feedback module connected with the main control processing module. The main control processing module constructs a three-dimensional grip curve surface model based on a pressure signal, calculates a blood vessel exposure index, dynamically adjusts a sensor weight according to patient classification, and forms a closed-loop control from signal collection to individualized intervention. The application reduces the workload of nurses through an automatic process, reduces the puncture failure rate through accurate grip evaluation, realizes accurate decompression guidance through clustering typing, and relieves psychological stress through multi-modal interaction. The standardized voice prompt and intelligent guidance reduce noise pollution in the blood collection area, improve the environmental comfort, and liberate nurses from repeated guidance, so that the nurses can pay more attention to core links such as sterile operation and improve the blood collection efficiency of the hospital.

Described herein are apparatuses and methods for classifying a patient as being asleep or awake. Such an apparatus can include an accelerometer and a processor. The accelerometer, alone or in combination with the processor, is used to determine an activity level of the patient and a posture of the patient. The processor is configured to classify the patient as being asleep in response to both (i) the posture of the patient being recumbent or reclined for at least a sleep latency duration, and (ii) the activity level of the patient not exceeding an activity threshold for at least the sleep latency duration; and classify the patient as being awake in response to at least one of (iii) the posture of the patient being upright for at least an awake latency duration, or (iv) the activity level of the patient exceeding the activity threshold for at least the awake latency duration.

The invention relates to the field of medical equipment based on the digital information technology, in particular to a nursing auxiliary system in the first-aid transfer process. According to the invention, big data is utilized to analyze the patient condition in the first-aid transfer process and provide auxiliary suggestions. A neural network model based on image processing is constructed, patient information is converted into an image in a specific format, patients are classified according to an image classification mode, and the system comprises an inquiry prompt module, an information acquisition module, an auxiliary diagnosis module, a first-aid database and a display unit. A trained neural network model used for image classification is built in the auxiliary diagnosis module, and after information of a current patient is mapped to a template image to obtain an information image of the current patient, the patient information image of the current patient is classified by using the trained neural network model. And finding the existing patient information closest to the classification of the patient information image of the current patient and the first-aid treatment information of the existing patient.

Methods and systems to train and use an ensemble of artificial intelligence / machine learning (AI / ML) models to extract information from social determinants of health (SDoH), including training each of multiple dimensionality reduction models to reduce dimensionality of socio-demographic variables associated with a respective one of multiple SDoH categories, training a predictive model to predict a patient behavior for a geographic region (e.g., risk of non-adherence to treatment regimens) based on dimensionally reduced SDoH (alone or in combination with selected socio-demographic variables and / or other data), training a patient classification model to classify patients based on prescription transactions, and / or training a regional similarity model to determine a measure of similarity between geographic regions based on SDoH and / or dimensionally reduced SDoH. Also disclosed are techniques to visually represent outputs of the models on a user-interactive display.

The invention discloses a blood glucose monitoring alarm reminding system, which belongs to the technical field of blood glucose monitoring and is characterized in that various blood glucose patients are subjected to reserve analysis to obtain various patient classifications, and corresponding dynamic condition models are set according to the patient classifications; analyzing the user, determining a patient classification corresponding to the user, configuring a corresponding blood glucose background recognition scheme for the user according to the patient classification, and configuring a dynamic condition model in a user side of the user; performing blood sugar monitoring on the user to obtain blood sugar monitoring data; the blood sugar background of the user is recognized in real time, and a blood sugar background label is marked on the blood sugar monitoring data according to the blood sugar background; performing real-time alarm analysis on the blood glucose of the user to obtain blood glucose monitoring data, and analyzing a blood glucose background corresponding to the blood glucose monitoring data through a dynamic condition model to obtain an alarm condition; analyzing the blood glucose monitoring data according to the alarm condition to obtain an alarm analysis result, and performing corresponding processing according to the alarm analysis result.

This invention relates to the field of medical data management technology, specifically to a method for managing stressors in hypertensive patients based on meticulous nursing care. After identifying blood pressure stress points, this method quantifies the conformity of short-term stress stimuli to identify both short-term and chronic blood pressure stress points. The impact of short-term blood pressure stress on the patient is then obtained through the rate of change in systolic blood pressure. Complete chronic stress phases are determined through splitting and merging methods, and the similarity between different time periods is analyzed to obtain the chronic blood pressure stress impact on each patient. Patients can be classified based on these two stress characteristics. In the results of the first classification, the similarity of the complete chronic stress phases is further considered for further classification. This invention determines the short-term and long-term blood pressure stress characteristics of patients based on the trends in systolic and diastolic blood pressure, enabling meticulous patient classification and facilitating effective management and targeted nursing care for hypertensive patients.

There is provided a method of classifying a biological status of an individual. The method comprising: obtaining a biological sample from a patient; obtaining health-related information from the patient, said information including patient gender; analysing the sample to identify a quantity of each of 2 or more endogenous analytes in the sample; comparing the analyte quantities to reference data from healthy individuals to classify the patient as healthy, pre-diseased, at risk of disease or diseased for at least one health-related condition. The reference data includes data derived from a group of biological samples of individuals having the same gender as the patient and not having a need for medical treatment for a disease or illness, each biological sample of the group of biological samples having been analysed by the same process as used to analyse the patient sample, the process being monitored to maintain a predetermined level of consistency.

A system for predicting and managing eye health risks in diabetes, consisting of a data acquisition interface for receiving ophthalmic images and longitudinal health data of a diabetes patient, a feature extraction module for deriving imaging and clinical features, a trained risk assessment engine for classifying the patient into risk or stable categories and generating alerts according to predefined thresholds, and a communication module for providing personalized recommendations and notifications to patient and physician interfaces.

The invention relates to a diabetes adjuvant therapy cloud platform system based on an artificial intelligence model, and relates to the technical field of health-related information systems and cloud platforms, and the diabetes adjuvant therapy cloud platform system comprises a correlation analysis module which is used for determining a plurality of patient classification factors based on medical records of a plurality of historical diabetic patients, determining a plurality of diabetic patient types and the complication risk of each diabetic patient type; the feature extraction module is used for determining the type of the diabetic patient of the current patient according to the medical record of the current patient and the plurality of patient classification factors, and extracting complication risk features from the medical record of the current patient; the auxiliary treatment module is used for determining similar historical diabetic patients according to the complication risk characteristics of the current patient and determining an auxiliary treatment scheme of the current patient based on the auxiliary treatment scheme of the similar historical diabetic patients and the complication risk characteristics of the current patient, and the diabetes diagnosis and treatment efficiency and the treatment effect are improved.

An information processing system 10 evaluates the worth of each of a plurality of inspection items. The information processing system 10 comprises: an information acquisition unit 201 that acquires patient information 215 on each of a plurality of patients to be inspected and inspection worth information 216 on an effect obtained by performing an inspection on each of the plurality of patients for the inspection items; a classification unit 202 that classifies the plurality of patients into a plurality of groups on the basis of the patient information 215 and the inspection worth information 216 that have been acquired by the information acquisition unit 201; and an inspection worth representative value calculation unit 204 that calculates, for each of the groups classified by the classification unit 202, an inspection worth representative value 218 that is an index value statistically obtained from the plurality of pieces of inspection worth information 216.

A method (100) for training a hemodynamic instability cause classification model to classify a patient's determined likelihood of future hemodynamic instability as caused by one or more of cardiogenic shock, septic shock, and hypovolemic shock, the method comprising: (i) receiving (120) a preliminary cause classification model training dataset; (ii) analyzing (130) the received dataset to classify each patient as likely to be hemodynamically unstable; (iii) analyzing (140) the received patient information by an outlier detection model to determine whether that patient's hemodynamic instability is likely to be cardiogenic shock, septic shock, hypovolemic shock, or an outlier; (iv) generating (150) a final cause classification model training dataset by removing each of the patients determined as being an outlier; and (v) training (160) the hemodynamic instability cause classification model to classify a cause of the patient's determined likelihood of future hemodynamic instability.

The invention discloses an epilepsy network classification method based on a graph attention network and an individual morphological brain network, and relates to the technical field of automatic identification and analysis of the individual morphological brain network. The method comprises two parts of contents: firstly, constructing an individual morphological brain network, including a T1MRI image preprocessing process, extracting nine morphological features to construct feature vectors, and constructing an individual morphological brain network matrix; secondly, constructing a graph attention network, wherein the graph attention network comprises a graph attention layer, a batch normalization layer, an activation function, a pooling layer and a full connection layer; the method can adaptively pay attention to the abnormal connection most related to the epilepsy disease, and effectively capture the complex relationship and feature information between the nodes in the brain region. Experimental results show that compared with a traditional individual morphology brain network identification and analysis method, the method has advantages in the aspects of brain region connection and topological attribute anomaly analysis and brain network-based patient classification, and more comprehensive and objective information can be provided.

The invention provides an AI blood sampling decompression wrist developer system and an application method. The system comprises a sensor matrix module; the main control processing module is connected with the sensor matrix module; the voice interaction module is connected with the main control processing module; the feedback module is connected with the main control processing module; and the main control processing module constructs a three-dimensional gripping force curved surface model based on the pressure signal, calculates a vascular exposure index, dynamically adjusts the weight of the sensor according to patient classification, and forms closed-loop control from signal acquisition to personalized intervention. The workload of nurses is reduced through an automatic process, the puncture failure rate is reduced through accurate grip assessment, accurate decompression guidance is achieved through clustering and typing, and psychological stress is relieved through multi-mode interaction. The noise pollution of a blood sampling area is reduced through standardized voice prompt and intelligent guidance, and the environmental comfort is improved; nurses are liberated from repeated guidance, more energy can be put into core links such as sterile operation, and the blood sampling efficiency of hospitals is improved.

The utility model discloses a medical record isolation signboard, which relates to the field of patient classification identification and comprises a signboard and three groups of shielding scrolls, the shielding scrolls are mounted inside the top end of the signboard and can be freely stretched, and the lower ends of the shielding scrolls are magnetically attracted on the inner wall of the bottom end of the signboard. The signboard comprises a plastic back plate, a first notice board, a second notice board, a third notice board, a magnetic attraction groove and a double-faced adhesive tape, a square groove position is formed in the front end of the plastic back plate, the medical record isolation signboard is divided into three types of surfaces, the three states of air isolation, droplet isolation and contact isolation can be displayed at the same time, and the medical record isolation signboard is convenient to use. When the signboard is used, two of the signboards can be shielded according to needs, only one state needing identification is left, the fixing mode of magnetic attraction is stable, the situation of falling is not prone to occurring, and the signboard is provided with the double-faced adhesive tape at the rear portion and can be pasted on the surface of a medical record of a patient so that medical staff can conveniently distinguish the signboard.

The invention discloses a specific disease intelligent follow-up visit management system based on multi-model fusion and a patient grading method, and belongs to the technical field of medical informatization. The system comprises a data integration layer, an intelligent patient classification and grading module, a model fusion layer, a personalized follow-up visit content generation module, a follow-up visit state management module, a data synchronization and integration module and a statistical analysis layer, high / medium / low risk grading is achieved, and the model fusion layer comprises recurrence, re-admission, bleeding and compliance prediction models. The intelligent patient classifying and grading module is used for classifying and judging based on rules of diagnosis sets, operation markers, bedridden duration, tumor markers and heart and cerebral vessel markers, and calculating a total score by adopting a category adaptive weight and a multi-dimensional score; patient grading accuracy is improved by more than or equal to 90%, VTE recurrence prediction accuracy is improved by more than or equal to 85%, recurrence rate is remarkably reduced by 27%, and follow-up visit efficiency is improved by more than 20%.

The application discloses a kind of inpatient hierarchical management method and system, belong to medical risk management technical field.The method includes: obtaining the personal information of inpatient and determining its risk index;According to risk index, dynamic correction electronic fence;Through real-time acquisition patient position by wearable device, combine fence relationship to calculate risk index;While calculating vital sign index according to vital sign information;Comprehensive risk index and vital sign index determine early warning grade, and execute corresponding early warning prompt.The application realizes the change from passive patrol to active early warning through multi-dimensional risk fusion and dynamic fence adjustment, improves the timeliness and accuracy of high-risk patient monitoring, while reducing false positive rate and nursing interference.

A real-time method for risk assessment and prediction of complications in plastic surgery. It includes a scoring method that classifies patients in distinct levels or risk-groups according to a risk score. According to the variables present in each individual, the algorithm estimates a risk factor for that particular data set provided by the patient through a questionnaire, calculates a risk score, classifies each patient in a risk group, and displays results in different electronic devices and personalized apps.

There is provided a method of classifying a biological status of an individual. The method comprising: obtaining a biological sample from a patient; obtaining health-related information from the patient, said information including patient gender; analysing the sample to identify a quantity of each of 2 or more endogenous analytes in the sample; comparing the analyte quantities to reference data from healthy individuals to classify the patient as healthy, pre-diseased, at risk of disease or diseased for at least one health-related condition. The reference data includes data derived from a group of biological samples of individuals having the same gender as the patient and not having a need for medical treatment for a disease or illness, each biological sample of the group of biological samples having been analysed by the same process as used to analyse the patient sample, the process being monitored to maintain a predetermined level of consistency.

The application discloses a patient large model classification method and system based on memory and retrieval enhancement, relates to the technical field of artificial intelligence, and inputs current patient information into a trained medical large model to output a classification result of the patient information; the training process of the medical large model is as follows: step one, different patient information is acquired to construct a training data set, and a corresponding knowledge instance of an i-th data in the training data set is obtained through a retrieval module; step two, each knowledge instance is subjected to a memory plug-in to obtain a knowledge instance representation; step three, the i-th data is input into a large language model, a gating mechanism is used to incorporate the knowledge instance representation corresponding to the i-th data into the information flow of the large language model, and thus the classification result of the i-th data is output; the patient large model classification method and system can dynamically adjust the used knowledge instance according to the specific condition of a patient, and thus the accuracy of patient classification is significantly improved.

The present invention relates to a device for obtaining a trained machine learning classification model for classifying a patient into at least two groups of subjects, wherein the subjects of said group have a gait pathology. The invention also relates to a device and method for classification of a patient in at least two groups of subjects, wherein the subjects of said group have a gait pathology.

The invention discloses a patient classification method and device based on magnetic resonance parameters, equipment and a medium. The method comprises the steps of obtaining target magnetic resonance parameters of a target patient; the target magnetic resonance parameter is screened from a plurality of preset magnetic resonance parameters based on the heart resynchronization tag of the sample patient and the real result of the occurrence of the preset endpoint event; the target magnetic resonance parameters comprise a myocardial strain parameter, a gadolinium contrast agent delay strengthening result and a left ventricular torsion parameter; inputting the target magnetic resonance parameters of the target patient into a target classification tree, and classifying the target patient by adopting the target classification tree to obtain a target classification result corresponding to the target patient; the target classification tree is obtained by extending a preset classification tree on the basis of cut-off values of the target magnetic resonance parameters and meeting the influence of the cut-off values on heart resynchronization; and outputting a target classification result. According to the method, the patients can be classified more accurately.