摘要
目的应用随机森林模型探讨慢阻肺患者报告结局(COPD-PRO)中各维度与modified Medical Research Council(mMRC)呼吸困难评分一致性评价的效果,以评估COPD-PRO对患者症状判断的准确性。方法于山西13所医院收集300例慢阻肺患者,以mMRC评分生成的二分类变量为结局变量,COPD-PRO13个维度为预测变量,并纳入人口学特征变量,构建随机森林模型,并与决策树模型进行比较。结果IND(独立性)、ANX(焦虑)、COG(疾病认知)、DEP(抑郁)等维度对于慢阻肺患者报告结局量表的症状评估的贡献较大。两种模型性能比较结果显示,随机森林的特异度、精准度和AUC和F 1值都高于决策树模型。结论随机森林模型在慢阻肺患者报告结局的症状评估中具有较好的预测效果,并识别影响患者症状的相关因素,为临床治疗与管理提供理论依据。
Objective To explore the effect of consistent evaluation of each dimension in chronic obstructive pulmonary patients-reported outcomes(COPD-PRO)and the modified Medical Research Council(mMRC)dyspnea score in patients by random forest model,and assess the accuracy of COPD-PRO in determining patients′symptoms.Methods Three hundred cases of patients with COPD were collected from 13 hospitals in Shanxi Province.Dichotomous variables generated with mMRC scores were used as the outcome variable,the 13 dimensions of COPD-PRO were used as the predictive variables,and demographic characteristic variables were included to construct a random forest model,which was compared with the decision tree model.Results IND(independence),ANX(anxiety),COG(disease cognition),SAT(satisfaction)and other dimensions contributed greatly to the symptom evaluation of COPD-PRO.The performance comparison results of the two models showed that the specificity,accuracy,AUC values and F1 values of random forest were higher than those of decision tree model.Conclusion The random forest model had a high application value in symptom assessment of COPD-PRO and identified relevant factors affecting patients′symptom,providing theoretical basis for clinical treatment and management.
作者
王莹
李莉芳
何航帜
杨青青
张垚烨
张岩波
赵卉
Wang Ying;Li Lifang;He Hangzhi(Department of Health Statistics,School of Public Health,Shanxi Medical University,Shanxi Provincial Key Laboratory of Major Disease Risk Assessment,(030001)Taiyuan)
出处
《中国卫生统计》
CSCD
北大核心
2023年第5期677-681,共5页
Chinese Journal of Health Statistics
基金
中央引领地方科技发展专项基金项目(YDZX20191400004736)
山西省重点研发计划项目(201903D421058)
国家自然科学基金项目(81872714)。
关键词
慢性阻塞性肺疾病
患者报告结局
随机森林
决策树
Chronic obstructive pulmonary disease
Patient-reported outcomes
Random forest
Decision tree
