Effects of meteorological factors on incidence of scarlet fever during different periods in different districts of China

To reveal the difference of meteorological effect on scarlet fever in Beijing and Hong Kong, China, during different periods among 2004–2014.The data of monthly incidence of scarlet fever and meteorological variables from 2004 to 2014 in Beijing and Hong Kong were collected from Chinese science data center of public health, meteorological data website and Hong Kong observatory website. The whole study period was separated into two periods by the outbreak year 2011 (Jan 2004–Dec 2010 and Jan 2011–Dec 2014). A generalized additive Poisson model was conducted to estimate the effect of meteorological variables on monthly incidence of scarlet fever during two periods in Beijing and Hong Kong, China.Incidence of scarlet fever in two districts were compared and found the average incidence during period of 2004–2010 were significantly different (Z=203.973, P<0.001) while average incidence became generally equal during 2011–2014 (Z=2.125, P>0.05). There was also significant difference in meteorological variables between Beijing and Hong Kong during whole study period, except air pressure (Z=0.165, P=0.869). After fitting GAM model, it could be found monthly mean temperature showed a negative effect (RR=0.962, 95%CI: 0.933, 0.992) on scarlet fever in Hong Kong during the period of 2004–2010. By comparison, for data in Beijing during the period of 2011–2014, the RRs of monthly mean temperature range growing 1°C and monthly sunshine duration growing 1h was equal to 1.196(1.022, 1.399) and 1.006(1.001, 1.012), respectively. The changes of meteorological effect on scarlet fever over time were not significant both in Beijing and Hong Kong.This study suggests that meteorological variables were important factors for incidence of scarlet fever during different period in Beijing and Hong Kong. It also support that some meteorological effects were opposite in different period although these differences might not completely statistically significant.