The Statistical Study Of The Association Between Air Pollution And Asthma Emergency Visits

Objectives Analyze the association between air pollution and asthma emergency visits, and discuss the potential risk factors of asthma. Thus we can provide useful information on air quality criteria suited to local places and meaningful information on the prevention of asthma, and in the meantime we can discuss the characteristic of the application of the time-series analysis and the case-crossover study in the study of the health effect of air pollution.Methods In the time-series analysis, we use correlogram to select the most ideal span, use locally weighted regression to treat the time trend in the time series, and then fit generalized additive models to get the relative risk of air pollution exposure. In the unidirectional retrospective case-crossover study and bidirectional case-crossover study, based on the theory of the case-control study, we choose the index period and the referent period, and then apply conditional logistic regression to get the odds ratio of air pollution exposure.Results When the day of lags is 0, the result of the single air pollutant model of generalized additive models indicates that PM₁₀ and fog are risk factors of athsma, while it's multiple air pollutant model shows that fog is also a risk factor of athsma. The most ideal generalized additive models selected by the step regression, i. e. model 1, shows that PM₁₀ is a risk factor of athsma. Model 2 is constructed by adding the independent variable fog to model 1. The outcome of model 2 indicates that fog is also a risk factor of athsma. The outcome of the single air pollutant model of the bidirectional case-crossover study shows that all of the air pollutants are not risk factors of athsma, whereas it's multiple air pollutant model shows that fog is a risk factor of athsma. The result of the single air pollutant model and multiple air pollutant model of the unidirectional case-crossover study shows that PM₁₀ is a risk factor of athsma. When the day of lags is 0, model 1 shows that every 44. 25μg/m³ rise in PM₁₀ concentration causes 6. 53% increase in asthma emergency visits. However, generally speaking, with the increase of lags, the relative risk of PM₁₀ is decreasing; Model 2 shows that every 44. 25μg/m³ rise in PM₁₀ concentration causes 5. 36% increase in asthma emergency visits. However, with the increase of lags, PM₁₀ is not risk factor of athsma any more. When the day of lags is 0, every one unit of increase in fog leads to 7. 7% increase in asthma emergency visits in model 2. With the increase of lags, the relative risk of fog is increasing. Minimum temperature and maximum temperature are not confounding factors.Conclusions The outcome of the time-series analysis is different from that of the case-crossover study. They both have merits and demerits. However, complicated statistical models are needed in the time-series analysis, and the requirement for softwares is also high. Therefore, the case-crossover study is simpler than the time-series analysis in practice.