Modelling distributed lag effects in mortality and air pollution studies: the case of Santiago

Most of the epidemiological literature on air pollution and mortality deals only with single or dual pollutant models whose results are hard to interpret and of questionable value from the policy perspective. In addition, much of the existing literature deals only with the very short-term effects of air pollution whereas policy makers need to know when, whether and to what extent pollution-induced increases in mortality counts are reversed. This involves modelling the infinite distributed lag effects of air pollution. Borrowing from econometrics this paper presents a method by which the infinite distributed lag effects can be estimated parsimoniously but plausibly estimated. The paper presents a time series study into the relationship between ambient levels of air pollution and daily mortality counts for Santiago employing this technique which confirms that the infinite lag effects are highly significant. It is also shown that day to day variations in NO2 concentrations and in the concentrations of both fine and coarse particulates are associated with short-term variations in death rates. These findings are made in the context of a model that simultaneously includes six different pollutants. Evidence is found pointing to the operation of a very short term harvesting effect.