基于主成分分析下慢性阻塞性肺疾病急性加重频率的判别分析

Discriminant analysis of acute exacerbation freque-ncy of chronic obstructive pulmonary disease based on principal component analysis

目的对慢性阻塞性肺疾病患者(COPD)一般资料及临床资料进行主成分分析和判别分析,研究影响疾病急性加重频率的因素,以预测患者后续发病频率。方法选取我院2016年1月~2017年10月收治的350例COPD患者为研究对象,收集患者例入院时一般资料、24h内动脉血气分析,首次降钙素原(PCT)含量、C-反应蛋白(CRP)含量及血常规,及入院结束治疗时的抗生素治疗时间、住院时间。对入选患者随访1年,并明确1年内慢性阻塞性肺疾病急性加重(AECOPD)次数,根据疾病加重次数将患者分为三组(A组:≤1次;B组:<3次;C组:≥3次)。比较三组一般资料及临床资料的差异,利用因子分析提取上述临床指标及一般资料主成分,并利用主成分对AECOPD患者急性加重频数行判别分析,比较提取主成分前后判别分析率差异。结果共纳入患者321例,其中A组105例、B组105例、C组111例,三组一般资料及临床资料除吸烟年无明显差异,其余指标组间比较,差异有统计学意义(P<0.05)。从临床指标及一般资料中共提取5个主成分,综合信息提取率83.3%。一般资料及临床指标判别分析提示,总体判别正确率为74.8%,A、B、C组三组正确判别率分别为88.6%、65.7%、70.3%;利用提取的主成分进行判别分析提示,总体判别正确率72.0%,A、B、C组三组正确判别率分别为97.1%、60%、59.5%。结论利用常见一般资料及临床资料进行判别分析,可预测患者发病频数,主成分分析的判别正确率与综合资料判别正确率间无明显差异。

Objective The factors influencing the frequency of acute exacerbation of chronic obstructive pulmonary disease(COPD)were studied by principal component analysis(PCA)and discriminant analysis in order to predict the frequency of acute exacerbation in patients with chronic obstructive pulmonary disease(COPD).Methods 350 patients with chronic obstructive pulmonary disease(COPD)admitted in our hospital from January 2016 to October 2017 were collected.Immediate general data including age,height,weight,medical history,smoking year and arterial blood gas analysis were collected.For the first time,content of procalcitonin(PCT)and the content of C-reactive protein(CRP)and blood routine were collected.The antibiotic treatment time and hospitalization time at the end of admission were collected.The patients were followed up for one year and the frequency of acute exacerbations of chronic obstructive pulmonary disease(AECOPD)was determined.According to the frequency of exacerbations,the patients were divided into group A(≤one time),group B(<three times)=and(group C≥three times).The differences of general data and clinical data among the three groups after admission were compared.Clinical indexes and general data were extracted by factor analysis.The frequency of acute exacerbation of AECOPD patients was analyzed by discriminant analysis using of above data and the extracted principal components.The difference of discriminant analysis rate before and after the extraction of principal components was compared.Results The general data and clinical data of the three groups had no statistical difference except smoking years,but there was statistical difference in the other indexes among the three groups(P<0.05).Five principal components were extracted from clinical indexes and general data,and the extraction rate of comprehensive information was 83.3%.The general data and clinical index discriminant analysis showed that the overall correct discriminant rate was 74.8%,and the correct discrimination rate of the three groups was 88.6%,65.7%and 70.3%.The discriminant analysis using the extracted principal components suggested that the correct rate of overall discrimination was 72%,and the correct discrimination rates of the 3 groups were 97.1%,60%,and 59.5%.Conclusion The patients with chronic obstructive pulmonary disease have different characteristics according to the different frequency of acute exacerbation.The principal representative components can be extracted by factor analysis from common clinical collected indexes.There is no significant difference between the discriminant accuracy rate of principal component analysis and the comprehensive data.