Logistic 函数模型在预测慢病患病率中的应用  

The application of Logistic model in predicting the prevalence of chronic non-communicable diseases

作　　者：严若华[1] 王杨[1] 李卫[1] 

机构地区：[1]北京协和医学院,中国医学科学院,国家心血管病中心,阜外心血管病医院,心血管疾病国家重点实验室国家心血管病中心,北京100037

出　　处：《中华疾病控制杂志》2014年第3期257-260,共4页Chinese Journal of Disease Control & Prevention

摘　　要：目的探讨不同Logistic函数模型在预测慢性非传染性疾病(慢病)患病率中的应用。方法利用我国建国以来4次大规模高血压抽样调查的数据,分别以年份、人均国内生产总值(gross domestic product,GDP)的对数和老龄化率(65岁以上人群所占比例)作为自变量,患病率作为因变量,并参照美国近50年来的高血压流行状况选取模型的上限,建立Logistic函数模型,对高血压未来的发展趋势进行预估。通过计算各模型的平均绝对误差(MAE)、均方误差(MSE)和决定系数(R2)比较其拟合效果。结果取40%为Logistic模型的患病率上界,以年份为自变量时,2010年的患病率估计值为20.35%,患病率将在2060年左右趋于稳定(MAE=0.735,MSE=0.704,R2=0.963);以人均GDP的对数为自变量时,预计2010年的患病率约为23.80%(MAE=0.896,MSE=0.969,R2=0.964);以老龄化率为自变量时,2010年的患病率将达到26.63%(MAE=1.004,MSE=1.659,R2=0.945)。结论 Logistic函数模型在理论上符合人们对未来疾病发展的预估,同时在实际中,可以找到其它国家和地区的疾病流行状况作为现实依托,预测结果较为可靠。其中以GDP和老龄化率作为自变量的模型更注重数据的临床意义。Objective To explore the application of Logistic model in predicting the prevalence of chronic noncommunicable diseases. Methods The prevalence of hypertension in China was used as an example. The year, logarithm of per capita gross domestic product（GDP） and aging rate（ the proportion of people over 65） were used respectively as independent variable, and prevalence rate as dependent variable, to build Logistic models. These models＇ fitting effects were compared by the mean absolute error（ MAE）, the mean square error（MSE） and the determination coefficient （R^2）. Results Taken 40% as the upper limit of the models, and the year as the independent variable, the predicted value of preva- lence in 2010 was 20. 35%, and the prevalence rate tended to stabilize in 2060 （MAE = 0. 735, MSE = 0. 704, R^2 = 0. 963 ） ; if x refers to ＂log（ per capita GDP）＂, the prevalence arrived to 23.80% in 2010 （MAE = 0. 896, MSE = 0. 969, R^2 =0. 964）; when it comes to ＂aging rate＂, the prevalence at the same time was estimated to be 26. 63% （MAE = 1. 004, MSE = 1. 659, R^2 = 0. 945）. Conclusions The Logistic models not only display a relatively fiat trend of prevalence in the future, but also have clinical significance in providing a reliable prediction results.

关 键 词：高血压 患病率 LOGISTIC模型 

分 类 号：R181[医药卫生—流行病学]