基于“三年化疫”理论探讨麻疹发病与前期气象因素的相关性并建立预测模型

目的：依据中医运气“三年化疫＂理论，探讨北京地区麻疹发病与2-3年前气象因素的相关性，并建立BP人工神经网络的医疗气象预测模型。方法：提取北京地区1970年-2004年35年的气象数据和麻疹发病数据，采用Spearman相关分析和BP人工神经网络的统计学方法，探讨麻疹发病与3年前及2年前气象因素的相关性，并建立麻疹发病的气象预测模型。结果：北京地区麻疹发病集中在初之气和二之气，春季为发病高峰期。相关分析显示，其发病与3年前或2年前的平均气温呈显著负相关（P＜0.05），与平均风速和平均相对湿度呈显著正相关性（P＜0.05）。神经网络结果显示，3年前或2年前的平均风速是影响麻疹发病最重要的气象因子；利用3年前的气象因子建模，其预测精度更优。结论：麻疹发病与3年前或2年前的气象因素之间均存在一定的相关性，并可建立气象预测模型，在一定程度上证明了"三年化疫"理论可用于麻疹发病的预测。

Objective: To study the correlation between the incidence of measles in Beijing area and meteorological factors in two or three years ago based on the theory of ‘pestilence occurring after three years’, and to establish artificial neural network of medical meteorological prediction model. Methods: The data of measles in Beijing area from 1970 to 2004 (35 years) and meteorological data at the same period were selected. The correlation between the incidence of measles and previous meteorological factors was explored and meteorological prediction model was established by using spearman correlation and back-propagation artificial neural network. Results: Measles was in a state of high prevalence during the period of 1st Qi and 2nd Qi, which from February to May, namely the spring. The incidence of measles was significantly negatively related to temperature in two or three years ago (P < 0.05), while was significantly positively related to wind speed and relative humidity in two or three years ago (P < 0.05). The result revealed that the average wind speed two or three years ago was the most vital factors of meteorological phenomena affecting the incidence of measles, and the predictive result was better than the other one when used the meteorological factors modeling in three years ago. Conclusion: The incidence of measles was correlated to meteorological factors in two or three years ago, and established meteorological forecasting model. To some extent, it proved that the theory of ’pestilence occurring after three years’ can be used to predict the incidence of measles.