Logistic回归模型拟合SARS发病及流行特征

Fitness of morbidity and discussion of epidemic characteristics of SARS based on logistic models

目的 利用Logistic回归模型预测SARS发病及流行趋势 ,研究SARS发病与流行特征及相应预防措施对疾病流行的影响 ,为预防与控制SARS提供科学依据。方法 使用Logistic回归模型拟合广东省、香港特别行政区、北京市及山西省累计发病数据 ,并绘制出SARS流行曲线。结果 对上述地区发病人数进行拟合 ,各地曲线拟合的决定系数R2 均大于 0 99,预测值与实际值符合 ,无显著性差异。从SARS流行曲线推算出各地发病高峰及时间分布 ,预防措施对疾病流行曲线有明显影响。结论 Logistic回归模型对SARS流行的拟合效果好 。

Objective To investigate the epidemic characteristics and trend of SARS and study these characteristics and corresponding prevention measures to the influence of the epidemic of SARS and provide evidence for the prevention and control by using logistic models.Methods Logistic model was used to fit the cumulative incidence of Guangzhou,HongKong,Beijing and Shanxi.Incidence curve was drawn from the models.Results The models got good fitting results,all determination coefficients are more than 0.99 .The predicted values accord with the true values well and there is no statistical significance between them.The time distribution can be calculated and the influence of preventive measures are discussed.Conclusion Logistic models got very good fitness results and can be used to study the influence of preventive measures to the occurrence and epidemic characteristics of SARS in theory.