基于“三年化疫”理论探讨香港地区痢疾发病与气象因素的相关性

目的：基于中医“三年化疫”理论，研究香港地区痢疾发病与当年及1～3年前气象变化的关联性，并建立气象-痢疾发病预测模型。

方法：利用香港1997—2019年共23年的痢疾发病数据和1993—2019年共27年的气象资料数据，通过BP人工神经网络方法，从不同时间维度探讨痢疾与气象的相关性，比较得出最佳时段的气象预测模型。

结果：痢疾的发病高峰在每年的四之气（7—8月），利用当年至3年前（共4年）的综合气象因素建立的预测模型效果最佳，精度达到70%；其中，重要的气象因素依次为当年的平均气压、3年前的平均相对湿度及平均太阳总辐射。

结论：香港地区的痢疾发病与当年及3年前的气象变化具有相关性，“三年化疫”理论有助于建立痢疾发病的早期预警模型。

Objective: To study the correlation between the incidence of dysentery in Hong Kong and meteorological factors in one 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 correlation between the dysentery data (from 1997 to 2019) and meteorological factors (from 1993 to 2019) in Hong Kong was analysied based on BP artificial neural network, and the best meteological-dysentery prediction model was identified across different periods.

Results: The peak of dysentery incidence was appeared in the 4th qi (e.g., July to August). The prediction model using the comprehensive meteorological factors from the current year to three years ago (four years in total) was the best one with an accuracy rate of 70%. The important meteorological factors included the average vapor pressure from the current year, the average relative humidity and global solar radiation from three years ago.

Conclusion: The incidence of dysentery in Hong Kong is associated with the meteorological changes of the current year and 3 years ago. The theory of ‘pestilence occurring after three years’ is useful to establish the early warning model of dysentery prediction.