Establishment and Validation of a Predictive Model for Hemorrhage

Summary

Participation Requirements

Description

This study has two main parts. The first part is to verify and optimize the clinical application effect of the existing prediction model. The clinical data of the acute ischemic stroke intravenous thrombolytic population is collected retrospectively, mainly including baseline indicators and 7 days a...

This study has two main parts. The first part is to verify and optimize the clinical application effect of the existing prediction model. The clinical data of the acute ischemic stroke intravenous thrombolytic population is collected retrospectively, mainly including baseline indicators and 7 days after thrombolysis Internal bleeding, based on the existing prediction models (HAT, SIT-sICH, THRIVE), calculate the prediction probability, and compare it with the actual bleeding situation, evaluate the clinical application effect of the prediction model, use ROC curve, calibration curve, sensitivity and Evaluation of indicators such as specificity. Using retrospective data, using multivariate logistic regression to analyze the predictive value of baseline clinical indicators, screening risk factors, and optimizing the HAT, SIT-sICH, and THRIVE prediction models. The logistic regression model is used to construct an improved HT prediction model based on the AIC principle; the method of model comparison is used to combine the clinical significance of the indicators to complete the construction of the prediction model. The second part is to evaluate the clinical application effect of the improved prediction model, and prospectively collect clinical data of AIS patients undergoing intravenous thrombolysis in Shenzhen Second People's Hospital, Shenzhen Longhua District People's Hospital, including general demographic data and laboratory tests Baseline indicators such as imaging examinations, bleeding within 7 days after thrombolysis, etc., were used to verify the improved HT prediction model using ROC curve, calibration curve, sensitivity and specificity, and external verification was performed to evaluate the prediction effect of the model.

Tracking Information