Intelligent critical risk prediction system for emergency department admitted patients

Abstract

This invention provides an intelligent critical care risk prediction system for emergency inpatients, comprising: a database module for storing a patient data sheet, which records at least one non-traumatic patient's information, including a medical record number, patient information, triage information, vital sign information, and chief complaint information; an analysis module for making predictions based on the non-traumatic patient's information to generate a ward transfer recommendation; and a processing module for receiving a query number via a human-machine interface, retrieving non-traumatic patient information with the same medical record number as the query number from the database module, inputting the patient information, triage information, vital sign information, and chief complaint information from the non-traumatic patient's information into the analysis module, and controlling the human-machine interface to display the ward transfer recommendation. This system can assist physicians in making ward decisions for patients.

Claims

1. An intelligent critical illness risk prediction system for emergency inpatients, applicable to an electronic device, the electronic device including a human-computer interface, the intelligent critical illness risk prediction system for emergency inpatients comprising: A database module stores a patient data table containing information on at least one non-traumatic patient. This non-traumatic patient data includes a medical record number, patient information, triage information, vital sign information, and chief complaint information, including age. An analysis module is generated by training multiple machine learning models based on the highest scores of the top N severity parameters. These severity parameters are extracted from the patient information, triage information, vital sign information, and chief complaint information. The vital sign information includes the first, last, and absolute values ​​of the last minus the first values ​​for body temperature, pulse, respiration, systolic blood pressure, and diastolic blood pressure. The analysis module makes predictions to generate a ward transfer recommendation. The ward recommendation is to transfer to a general ward or to an intensive care unit, where N is a natural number greater than 1; and a processing module electrically connected to the database module and the analysis module, which receives a query number from the human-machine interface to retrieve non-traumatic patient data with the same medical record number as the query number from at least one non-traumatic patient data stored in the database module. The processing module inputs patient information, triage information, vital signs information and chief complaint information from the non-traumatic patient data into the analysis module and controls the human-machine interface to display the ward transfer recommendation.

2. The intelligent critical care risk prediction system for emergency inpatients as described in claim 1 further includes a feature extraction module, a feature selection module, and a training module. The feature extraction module is electrically connected to the database module and is used to extract features based on patient information, triage information, vital sign information, and chief complaint information of at least one non-traumatic patient to generate severity parameter features. The feature selection module is electrically connected to the feature extraction module and is used to analyze the severity parameter features based on a feature selection algorithm to obtain a score ranking of the severity parameter features. The training module is electrically connected to the feature selection module and is used to train multiple machine learning models based on the top N highest scores among the severity parameter features to generate the analysis module.

3. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 2, wherein, The feature extraction module employs either principal component analysis or linear discriminant analysis.

4. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 2, wherein, The feature selection algorithm employs a filtering method, a wrapping method, or an embedding method.

5. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 2, wherein, The analysis module is generated by concatenating at least three of the trained machine learning models.

6. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 5, wherein, The multiple machine learning models include a linear regression model, a random forest model, a lightweight gradient boosting machine model, a support vector machine model, a K-nearest neighbor algorithm model, an XGboost model, a GRU model, a transformer model, a recurrent neural network model, and a long short-term memory model.

7. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 2, wherein, The feature selection module also includes the following parameters related to the severity of the illness: emergency observation time, admission method, the last stroke minus the first stroke of the total consciousness score, the first stroke of the total consciousness score, emergency medical treatment time, the last stroke of the total consciousness score, final triage level, and data on the cardiovascular and nervous systems.

8. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 1, wherein, The processing module performs missing value processing on the patient information, triage information, vital sign information, and chief complaint information by treating the missing values ​​as null values, thereby performing preprocessing on the patient information, triage information, vital sign information, and chief complaint information.

9. The intelligent critical illness risk prediction system for emergency inpatients as described in claim 1, wherein, The human-computer interface also displays several decision options for selection, namely, transferring to a general ward and transferring to an intensive care unit.

Images