strong seasonal patterns in daily mortality. The replication
of these findings in a city with a very different seasonal
pattern, and no cold season, is notable. Specifically,
an association between mortality and air pollution exists
in a location where there is very little likelihood of confounding
due to season or epidemics related to cold
weather. In addition, the apparently stronger relative impact
of air pollution on a population residing in a lessdeveloped
country is plausible. Residents of Bangkok are
more likely to spend greater time outdoors and are less
likely to have air conditioning than residents in areas
where most previous air pollution mortality studies have
been conducted. Therefore, it is likely that the effective
dose of any given concentration measured from a fixed
site outdoor monitor is also greater than that observed in
the Western industrialized countries. As indicated by Table
5, when the more complete and accurate data set was utilized
(i.e., with estimated PM10 for the first two years), the
RRs associated with natural, cardiovascular, respiratory,
and elderly mortality were often greater than those observed
in previous studies undertaken in the West. Cardiovascular
mortality exhibited a less clear-cut association,
and was lag-dependent. This may be due to errors in
cause-of-death coding, use of hospitals and medical care
in Bangkok, or simply chance.