Research On The Relationship Of Trauma Emergency With Weather Conditions And Its Forecasting Models In Beijing City

Medical meteorology, as a new cross discipline, aims to explore and explain the causal link between weather, climate and disease,health. With the intensification of global climate anomalies, the impact of meteorological conditions on human health has attracted increasing attention. In common diseases, trauma ranks the fourth cause of death in our country, with the characteristics of strong burst, complex and diverse, and treatment difficulty, after malignant tumor, cardiovascular and respiratory disease. So, for "injury prevention, health promotion", to strengthen the basic research of trauma, looking for high-risk meteorological elements, reducing the incidence of injury, has important practical value. Based on the research progress at home and abroad, all trauma emergency cases during 2008 to 2012,from three comprehensive first class hospitals of Beijing city and the meteorological data of Beijing in the same period are collected. Firstly, on the basis of description of epidemiological features of trauma emergency (population characteristics, injured parts, trauma classification) of Beijing city, the correlation between trauma (total trauma, dog bites and traffic injury) emergency number and meteorological elements is analyzed. Secondly, mining three types of trauma emergency number’s relationships with the weather conditions (weather types), through weather typing, comparative statistics and case analysis. Thirdly, by using generalized additive models and eliminating confounding factors, to quantitative evaluate the exposure response relationship between daily average trauma emergency number and meteorological factors, build multivariate stepwise regression prediction models and generalized additive prediction models of trauma number, and test the grade prediction effect, providing the weather forecast service for public having a safe travel. The main results are as follows:(1) The total number of trauma cases from three comprehensive hospitals in Beijing city during 2008-2012 is 195826, exception of unexplained trauma, dog bites and traffic injury patients mostly. Trauma occurs in 1～90 years old. The age of high incidence of total trauma is 18～55 years old, with 18～35 years old mostly, of dog bites are 1～10,20～30 and 50～60 years old, of traffic injury is 18～60 years old. Head face, upper limb and lower limb are the major injury sites, with the proportion 31.34%, 28.11% and 22.49% respectively, accounting for 81.94% of the ratio. For Total trauma and traffic injury, the patients in male are more than female, with 1.3:1 and 1.5:1 of men and women ratio.For dog bites, the patients in female are more, with 0.8:1 of the ratio.(2) The total number of trauma is overal increasing with year. June, July and August in summer is the peak for total trauma, dog bite clinic, the lowest in February or December, with obvious effect on the weekend. For traffic injury, may is the peak month, with significant Friday effect and May Holiday effect. The emergency numbers of total trauma,dog bites and traffic injury all have the most significant positive correlations with the average temperature of (r=0.734,0.709,0.104,p<0.01), the second is dog bites showing a significant positive correlation with relative humidity (r=0.374,p<0.01), traffic injury showing a significant positive correlation with sunshine hours (r=0.065,p<0.01).(3) Different seasons, trauma happened in different weather types. In summer, high temperature, heavy rainstorm or the greatly temperature mutation can lead to the daily emergency number of traffic injury and dog bites increased substantially, while heavy rain and humid weather could decline the emergency number.The high weather type of total trauma and dog bites are fine and hot weather in spring and cool weather type in autumn, with the high incidence of traffic injury in the spring fine and hot weather, and in autumn cold weather type.(4) Total trauma and the temperature varies roughly linearly, especially in spring and autumn, with the most significant linear relationship (P<0.01)during -5～25℃of average temperature, total trauma patients increased by 1.73% for each increased 1℃.In the dead of winter or high summer, when the average temperature lower than-5.0℃ or higher than 25℃,the significance of linear relationship was lower (P<0.1), for each increased 1℃, average daily emergency number increased 0.94% and 0.05%.The exposure response curve of total trauma with relative humidity is inverse U-shape, when below 76% in the autumn, trauma patients increased by 0.42% for each increased 1%, while decreased by 0.59% when higher than 76% in summer.Compared with total trauma, the emergency number of dog bites has the higher correlation with the average temperature, when higher than 5.6℃,increased by 2.46% for each increased 1℃. While the relative humidity followed, increased by 0.58% for each increased 1%, when less than 72%. For traffic injury, the quantitative influence of meteorological elements is relatively weak, emergency number increased by 0.67% for each increased 1℃, while reduced by 0.16% for each increased 1%.(5) The prediction effect of Multiple stepwise regression and GAM for the total trauma emergency number is good, with the forecast accuracy of 89.5% and 90.0%, while the forecast accuracy of trauma risk levels were 78.6% and 76.8%.For the prediction of dog bite emergency number, the forecast accuracy of stepwise regression modeling is 99.7%, which higher than the forecast accuracy of GAM (72.3%).The predictability of traffic injury is relatively low, with only 42.7% forecast accuracy and 38.3% forecast accuracy of risk levels.In short, the high-risk group of trauma emergency in Beijing city, is the 18～35 years old young man, with the peak in summer of July, with the most significant correlation with temperature and relative humidity. The female of dog bites is more, and the preschool children less than 10 years old are vulnerable groups. Traffic injury peak in May. High temperature and heavy rainstorm could significantly increase the risk of trauma. Constructing the forecasting model of trauma emergency number through time series and meteorological elements is good, which has important reference value to realize the early warning and prevention of trauma.