手足口病预后因素的因子分析及预后预测模型

目的 了解手足口病的预后因素并建立预后预测模型。方法 收集2009—2016年在郑州市某医院住院的478例手足口病资料,用因子分析提取预后因素的公因子,用logistic回归建立基于因子得分的预后预测模型。结果 共提取5个公因子,累计方差贡献为86.65%。因子1包括外周血中性粒细胞计数和淋巴细胞比例,因子2包括外周血白细胞计数和血小板计数,因子3包括收缩压和发热持续时间,因子4为C反应蛋白,因子5为空腹血糖。预后预测模型为P＝1/1＋e^{-(0.0673X1+0.3126X2+0.0136X3+0.6502X4+0.2083X5+0.4080X6-0.3302X7+0.4257X8-3.043)}]。结论 基于因子分析的logistic回归模型能较好预测手足口病预后,应关注预后因素,以便做好预警和及时治疗工作。

Factor analysis of prognosis factors for hand,foot and mouth disease and predictive model for its prognosis

Objective To explore the prognosis factors of hand, foot and mouth disease (HFMD) and establish the predictive model of prognosis. Methods We collected the data about 478 HFMD cases hospitalized in a hospital in Zhengzhou city from 2009 to 2016. Factor analysis method was used for extracting common factors of prognosis factors, and logistic regression forestablishing predictive model of prognosis based on factor scores. Results There were 5 common factors extracted, contributing 86.65% of the total variance. The first factor contained peripheral blood neutrophil count (X2) and lymphocyte ratio (X7), the second factor peripheral blood white cell (X1) and platelet counts (X3), the third factor systolic pressure (X5) and fever duration (X8), the fourth factor C-reactive protein (X6) and the fifth factor fasting blood glucose (X4). The predictive model of prognosis wasP=1/1＋e^{-(0.0673X1+0.3126X2+0.0136X3+0.6502X4+0.2083X5+0.4080X6-0.3302X7+0.4257X8-3.043)}]. Conclusions Logistic regression model based on factor analysis can well predict the prognosis of HFMD. It is necessary to pay attention to prognosis factors so as to better implement early warning and perform timely treatment.