Logistic回归模型鉴别缺铁性贫血和地中海贫血的研究

Differential Diagnostic Study of Iron Deficiency Anemia and Thalassemia Based on Logistic Regression Modeling

目的:建立Logistic回归模型,用于鉴别缺铁性贫血和地中海贫血,并评价和验证模型的有效性。方法:收集2021~2022年昆明医科大学第一附属医院确诊缺铁性贫血和地中海贫血各121例患者资料,随机每组抽取100个病例样本数据为实验组,剩余21个样本数据为验证组。经单因素分析,确定具有统计学差异的变量,并将其纳入多因素分析,可得性别、Hb、MCHC和RDW⁃SD 4个变量为独立影响因素,并用这4个变量建立Logistic回归模型,以受试者操作特征(ROC)曲线下面积、灵敏度、特异度三个指标评估模型鉴别诊断效果。结果:单因素变量分析显示,RBC、Hb、HCT、MCV、MCH和MCHC值,缺铁性贫血组均低于地中海贫血组,RDW⁃SD值缺铁性贫血组高于地中海贫血组,组间差异具有统计学意义(P<0.05)。将性别、Hb、MCH、MCHC和RDW⁃SD带入二元Logistic回归分析,结果显示,RDW⁃SD值越高患缺铁性贫血的风险越高,Hb和MCHC下降越明显患缺铁性贫血的风险越高(P<0.05)。Logistic回归模型为:Logit(P)=-19.288-1.685X1+0.035X2+0.066X3-0.076X4。实验组的Logistic模型ROC曲线下面积为0.947,灵敏度为90%,特异度为91%;验证组的灵敏度和特异度分别为100%和72.4%。结论:基于Logistic回归的数学模型在缺铁性贫血和地中海贫血的鉴别诊断中有一定临床应用价值。

Objective:The objective of the research is to establish a logistic regression model for the differential diagnosis of iron deficiency anemia(IDA)and thalassemia(TT)and to explore its application value.Methods:The research analyzed medical records of 121 patients diagnosed with both IDA and TT at the First Affiliated Hospital of Kunming Medical University between 2021 and 2022.A total of 100 specimens were arbitrarily chosen from each cohort to constitute the experimental groups,while the remaining 21 were designated for validation reasons.After performing a univariate comparison to establish statistical significance,the subsequent action was to identify variables with significant differences for the multi⁃factor analysis.Based on the analysis,four variables⁃gender,hemoglobin,mean corpuscular hemoglobin concentration,and red cell distribution width⁃were identified as independent factors that influence the outcome.These four variables established a logistic regression model,which was evaluated for its differential diagnosis effect by analyzing sensitivity,specificity,and other indicators on a subject’s working characteristics curve.Results:The results of the univariate comparison illustrate that the IDA group has lower values of red blood cells,hemoglobin,hematocrit,mean corpuscular volume,mean corpuscular hemoglobin,and mean corpuscular hemoglobin concentration in their peripheral blood compared to the TT group.Conversely,higher values of peripheral blood red cell distribution width⁃standard deviation are observed in the IDA group in comparison to the TT group,and the difference is statistically significant with a P⁃value less than 0.05.Gender,hemoglobin(Hb),mean corpuscular hemoglobin(MCH),mean corpuscular hemoglobin concentration(MCHC),and red cell distribution width(RDW⁃SD)were analyzed via binary non⁃conditional logistic regression.The study showed that higher RDW⁃SD values were linked to an increased risk of IDA,while more significant reductions in Hb and MCHC were associated with a higher risk of IDA at a significance level of P<0.05.The logistic regression model is presented as:Logit(P)=-19.288-1.685X1+0.035X2+0.066X3-0.076X4.In the experimental group,the area under the receiver operating characteristic(ROC)curve of the logistic model is 0.947,the sensitivity is 90%,and the specificity is 91%;the sensitivity and specificity of the verification group are 100%and 72.4%,respectively.Conclusions:Logistic regression models have certain application value in the differential diagnosis of IDA and TT.