自发性深部脑出血预后预测模型的开发与内部验证

Development and Internal Validation of Prediction Model for Prognosis of Spontaneous Deep Intracerebral Hemorrhage

目的:探讨自发性深部脑出血(SDICH)患者预后不良的独立预测因素,建立预测模型并进行内部验证。方法:回顾性收集2019年1月—2022年4月台江县人民医院收治的423例SDICH患者的临床资料并进行分析,建立SDICH患者预后预测模型,并以6∶4随机拆分数据进行内部验证。结果:预测模型推导队列的受试者工作特征曲线下面积(AUC)为0.889,验证队列为0.912。验证队列中Hoslem-Test(P=0.84>0.05),Brier=0.077,预测模型验证队列GiViTI校准曲线带的95%CI区域均未穿过45°对角平分线,说明预测模型的预测概率与实际观测概率接近,具有较强的一致性。结论:本研究所建立的预测模型预测SDICH患者治疗3个月后预后不良的准确性高,有助于提高此类患者的早期识别和筛选能力。

Objective:To investigate the independent predictors of poor prognosis in patients with spontaneous deep intracerebral hemorrhage(SDICH),establish a prediction model and conduct internal validation.Methods:The clinical data of 423 patients with SDICH in Taijiang County People's Hospital from January 2019 to April 2022 were retrospectively collected and analyzed.A prediction model for prognosis of SDICH patients was established.The data were randomly split at a ratio of 6:4 for internal validation.Results:The area under receiver operating characteristic curve(AUC)of the prediction model derivation cohort was 0.889,and that of the validation cohort was 0.912.In the validation cohort,Hoslem-Test was performed(P=0.84>0.05)and Brier=0.077.The 95%CI region of calibration curve band of the prediction model validation cohort did not cross the 45°diagonal bisector,indicating that the prediction probability of the prediction model was close to the actual observation probability,with strong consistency.Conclusion:The prediction model established in this study has a high accuracy in predicting the poor prognosis of SDICH patients after 3-month treatment,which is helpful to improve the early identification and screening ability of such patients.