基于Fine-Gray竞争风险模型预测老年肺结核结核特异死亡的列线图的构建  

Constructing a nomogram for predicting tuberculosis-specific mortality in elderly with pulmonary tuberculosis based on Fine-Gray competitive risk model

作　　者：谷若画 吕全军 任鹏飞[3] 王重建[4] 毛振兴 霍文倩 牛恺琳 陈裕[3] 李琳琳[1] GU Ruohua;LYU Quanjun;REN Pengfei;WANG Chongjian;MAO Zhenxing;HUO Wenqian;NIU Kailin;CHEN Yu;LI Linlin(Department of Health Statistics,College of Public Health,Zhengzhou University,Zhengzhou 450001;Department of Nutrition,the First Affiliated Hospital,Zhengzhou University,Zhengzhou 450052;Department of Tuberculosis,the Sixth People′s Hospital of Zhengzhou,Zhengzhou 450015;Department of Epidemiology,College of Public Health,Zhengzhou University,Zhengzhou 450001;Department of Environmental Health,College of Public Health,Zhengzhou University,Zhengzhou 450001)

机构地区：[1]郑州大学公共卫生学院卫生统计学教研室,郑州450001 [2]郑州大学第一附属医院营养科,郑州450052 [3]郑州市第六人民医院结核病科,郑州450015 [4]郑州大学公共卫生学院流行病学教研室,郑州450001 [5]郑州大学公共卫生学院环境卫生学教研室,郑州450001

出　　处：《郑州大学学报（医学版）》2023年第3期367-372,共6页Journal of Zhengzhou University(Medical Sciences)

基　　金：河南省重点研发与推广专项(科技攻关)(212102310195);河南省医学科技攻关计划项目(201702328)。

摘　　要：目的:建立预测老年肺结核结核特异性死亡的列线图模型。方法:选取2015年1月至2021年2月在河南省14家传染病医院确诊肺结核的544例老年(≥65岁)患者,随机分为训练集(n=435)和验证集(n=109)。在训练集中,利用累积发生函数(CIF)评估患者结核特异性死亡3 a和5 a的累积发生率。以结核特异性死亡为研究终点,采用Fine-Gray竞争风险模型筛选预后预测因素。绘制列线图,采用一致性指数(C指数)和校准曲线在训练集与验证集数据中评估模型预测性能。结果:训练集数据中结核特异性死亡3 a和5 a的累积发生率分别为9.0%、9.7%。Fine-Gray竞争风险模型分析结果显示年龄、治疗类型、胸片空洞和血清白蛋白水平是老年肺结核患者结核特异性死亡的预测因素,由此构建的列线图在训练集和验证集中,预测的C指数(95%CI)分别为0.815(0.798~0.832)和0.851(0.805~0.897),校准曲线显示预测的3 a和5 a死亡率与实际死亡率有着较好的一致性。结论:Fine-Gray竞争风险模型能有效识别老年肺结核结核特异性死亡的影响因素,以此为依据构建的列线图模型能有效预测患者预后。Aim:To construct a nomogram model for predicting tuberculosis-specific mortality in elderly with pulmonary tuberculosis(PTB).Methods:A total of 544 elderly patients with PTB who were hospitalized in 14 infectious disease hospitals in Henan Province from January 2015 to February 2021 were selected and randomly allocated into training cohort(n=435)and validation cohort(n=109).In the training cohort,cumulative incidence function(CIF)was used to evaluate the 3-year and 5-year tuberculosis-specific mortality.Fine-Gray competitive risk model was used to screen the prognostic predictors,and a nomogram was drawn.The concordance index(C-index)and calibration curve were used to evaluate the predictive accuracy of the nomogram model based on the data of training cohort and validation cohort.Results:In the training cohort,the 3-year and 5-year CIF of tuberculosis-specific mortality were 9.0%and 9.7%,respectively.Fine-Gray competitive risk model showed that age,treatment type,cavities on chest X-rays,and serum albumin level were predictors of tuberculosis-specific mortality.The C-index(95%CI)of the constructed nomogram was 0.815(0.798-0.832)and 0.851(0.805-0.897),respectively.The calibration curve showed that the predicted 3-year and 5-year tuberculosis-specific mortality were in good agreement with the actual mortality.Conclusion:Fine-Gray competitive risk model can effectively identify the prognostic factors of elderly patients with PTB,and the nomogram prognostic model based on this basis can effectively predict the prognosis of patients.

关 键 词：肺结核 结核特异性死亡 竞争风险模型 列线图 

分 类 号：R521[医药卫生—内科学]