||Statistical analysis of acute health effects of air pollution
||Andersen, Klaus Kaae (Mathematical Statistics, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
||Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
||This thesis deals with statistical methods and the application of these on the association
between ambient air pollution in Copenhagen during the years 2003-2009 and stroke
incidents occurring in the city. The purpose of the study is to investigate whether the
exposure to air pollution is a risk factor for experiencing a stroke.
Air pollutants treated are particulate matter measured as PM10, PM2.5, and PM7−800nm,
and the traffic pollutant NO2. Air pollution data is given as an average of measurements
from three monitors located in downtown Copenhagen, and the measurements are daily
averages. Stroke admission data considered is information about 8808 patients with
ischemic stroke and without atrial fibrillation (the most common type of arrhythmia.
An arrhythmia is a problem with the rate or rhythm of the heartbeat).
In order to explore the association between ambient air pollution and the number of
acute hospitalised stroke patients, a systematically working process is needed. Therefore,
the performing process is divided in several steps. First, descriptive introductions
of the stroke and air pollution data are given. Second, an exploratory data analysis
is performed and a theoretical introduction to the statistical approach is given. Two
methods have been applied for comparison. Both a time series analysis and a casecrossover
analysis have been considered using generalised additive models and conditional
logistic regression respectively. Furthermore two techniques have been applied
to the pollutants in each case, namely moving average and unconstrained distributed
lag. Third, the results from both the time series analysis and case-crossover analysis
are presented in two separate chapters. Each chapter includes discussions of the results
and a conclusion. Last, an overall discussion and a conclusion is given together with
thoughts and considerations about future work.
The results from the analysis using both generalised additive models and conditional
logistic regression showed somewhat the same results. Particulate matter was found
not to be significant associated with stroke incidents, although the ultrafine particles,
PM7−800nm, showed associations for certain subgroups of the stroke population. NO2
was found significantly associated with stroke incidents for all patients as well as for
subgroups without certain co-morbidity or unhealthy habits. It appears from these
finding that traffic is the main contributor for the relation between stroke incidents and
ambient air pollution in Copenhagen.
||Technical University of Denmark (DTU) : Kgs. Lyngby, Denmark
Creation date: 2010-06-18
Update date: 2010-06-18