Introduction
There are many factors that caused the bad smog phenomene in Beijing. For example the automobile usage has been increase in Beijing since 1978, and till 2013 there are over 5 million automobiles registered; therefore, increase vehicle exhaust emissions contribute a good portion to PM2.5 readings. Other important factor is the population, Beijing is the capital of China; therefore, it has better employment opportunities, medical treatment, and education qualites compare to other cities. Till 2013, the permanent residents in Beijing have exceed 20 million, and along with the increase in residents, the cost of energy use also increase dramatically. The major pollution sources in Beijing comes from coal emissions, fuel emissions, and natural gas emissions, which are all connected to population.
The PM2.5 caused by coal emissions breakdown into three components, first is the particulate matter from the smoke of coal, second are the surfer oxide and nitrogen oxide that become secondary sulfate and nitrate through photochemistry, third is the incomplete combustion of volatile organic compounds from coal itself enters the atmosphere. Normally, the temperature decreases as the altitude increases, and create convection between low level and high level of atmosphere, which could spread pollutant towards higher altitude in order to reduce the pollution in the city. However, when a phenomenon called inversion occurs the result is just the opposite, atmospheric temperature increases as the altitude increases. The atmospheric structure above the city no longer acts as a filter, but a cap that traps the pollutant inside.
Objectives
Generalized-additive model, semiparametric Poisson regression model, and time-stratified case-crossover design will be compared in this paper to analyse the association between mortality and particulate air pollution.
Methods
In Xu Qin’s Fine Particulate Air Pollution and Hospital Emergency Room Visits for Respiratory Disease in Urban Areas in Beijing, data were collected from ten hospitals across six different district, and covers over ninety thousands people. A generalized-additive model with the Poisson link function was used to analyse the data, and the variables involved includes temperature, relative humidity, day of the week, public holidays, and influenza outbreaks(Qin et al., 4). The diseases focused in this study are respiratory disease, upper respiratory tract infection (URTI), lower respiratory tract infection (LRTI), and acute exacerbation of chronic obstructive pulmonary disease (AECOPD) (Qin et al., 6).
The result uses the daily patients’ emergency room visits (ERV) verse the PM2.5 concentration to show the association between air pollution and public health. In Susanne Breitner’s study Sub-micrometer particulate air pollution and cardiovascular mortality in Beijing, they select semiparametric Poisson regression model for data analyse, and the PM2.5 concentration they use are particle mass and surface area instead of numbers of particle. The diseases analysed in this paper are cardiovascular diseases, ischemic diseases, and stroke (Breitner et al., 5197).
The variables in this model are the same as in Qin’s. The data are presented as mortality varse the PM2.5 concentration, which is similar to Qin’s result.In Yuming Guo’s article The relationship between particulate air pollution and emergency hospital visits for hypertension in Beijing, uses time-stratified case-crossover design to connect the dot between air pollution and hypertension. Different from previous two cases, in this study each day of the week have three matching control days to produce a unbiased results due to confounding variables (Guo et al., 4447).
Discussion of results
The result from Qin’s article provide strong evidence for PM2.5’s effect on respiratory disease morbidity in Beijing. This result are more precise compare other similar studies because it collects data from six different district and 17 different hospital sites. However, due to confidential reasons, individual patient’s residencial information are not available for more accurate PM2.5 readings measurement. Breitner’s result showed that cardiovascular disease morbidity has increased due to short-term exposure to PM2.5 pollution, and other pollutant produced locally or transported from other area also have effects in Beijing.
The limitation in this study are inheritable measurement error that decrease the accuracy of the result. The time-stratified case-crossover design used to analyse the relationship between PM2.5 pollution and EMR for hypertension showed positive result, as levels of PM2.5 pollution increase EMR for hypertension also increase in Beijing. The limitation in this study is the lack of enough data collection because it only collects patients data from one hospital and PM2.5 data for one site; therefore, the result loses some precision and are not representative enough.
Conclusions
In conclusion, Xu Qin’s study include the most hospital, and air quality data; therefore it is the most accurate among the three studies. The final results presents a clear correlation between disease morbidity and PM2.5 concentration. Breitner’s methods not only test the particle number concentration, but surface area concentration and mass concentration as well. The particle SC and MC plays important role in determining the biological activity of particles(Brook, 2008; Valavanidis et al., 2008).
These additional methods helps determine the association in a biological precipitation. The case-crossover design from Yuming Guo’s article is a model that can control confounders that correlate to age, gender, and smoking habits, and as well as seasonal patterns. In other words, this design provides more detailed analysis, and gives impartial gauges in presence of strong seasonal confounding(Basu et al., 2005; Lee and Schwartz, 1999).
The disadvantage of the method in this paper is the diversity in each selected study. The first study test for respiratory disease, the second study test for cardiovascular disease, and the third tudy test for hypertension. Even though they all makes association with PM2.5 pollution, the conclude results might be too broad, and lacks representation. The advantage are that these studies provide a vast amount of models which presents more alternatives for future researchers. Overall, the methods used in each selected study has its own advantage and disadvantage.
