脑卒中危险因素分析及预测模型的构建

The analysis of risk factors of stroke and construction of prediction model

目的分析脑卒中的相关危险因素,并通过危险因素构建脑卒中风险预测模型。方法回顾2017年6月至2020年5月期间脑卒中和非脑卒中住院患者349例的病历资料,从中提取患者的年龄、性别等基本信息及高血压、糖尿病等疾病状况,从而分析脑卒中患者和非脑卒中患者的临床特征,通过LASSO分析筛选出潜在的危险因素,并在此基础上构建脑卒中预测模型。结果通过LASSO分析筛选出年龄、锻炼、高脂血症、高血压、房颤及糖尿病为最终预测因素,并通过该6个常见危险因素构建脑卒中的风险预测模型。模型的外部验证显示C指数为0.833(95%CI:0.749-0.917),该模型具有良好的准确性和校准能力,临床决策分析显示发生脑卒中阈值确定为>5%和<91%时,该临床预测模型是有意义的。结论本文分析了脑卒中的6个危险因素,构建了脑卒中风险预测模型,外部验证显示该预测模型具有较好的准度,医者可根据结果采取有针对性的预防策略,从而降低脑卒中的发病率。

Objective To analyze the related risk factors of stroke,and build a risk prediction model of stroke by risk factors.Methods The medical records of stroke and non—stroke hospitalized patients from June 2017 to May 2020 in Zhejiang Provincial People's Hospital were reviewed.A total of 349 patients were taken as the research objects,from which basic information such as the patient's age,gender,and disease states such as hypertension and diabetes were extracted.In order to analyze the clinical characteristics of stroke patients and non—stroke patients,the potential risk factors were screened by LASSO analysis,and a stroke prediction model was constructed on this basis.Results Through LASSO analysis,age,exercise,hyperlipidemia,hypertension,atrial fibrillation and diabetes were selected as the final predictive factors,and the six common risk factors were used to construct a risk prediction model for stroke.Through LASSO analysis,age,exercise,hyperlipidemia,hypertension,atrial fibrillation and diabetes were selected as the final predictive factors,and the six common risk factors were used to construct a risk prediction model for stroke.Conclusion This article analyzes the six risk factors for stroke and constructs a stroke risk prediction model.External verification shows that the prediction model has good accuracy,and doctors can take targeted prevention strategies based on the results,thereby reducing the incidence of stroke rate.