Quantcast

Case-crossover design: air pollution and health outcomes.

Research paper by Mieczysław M Szyszkowicz, Neil N Tremblay

Indexed on: 17 Aug '11Published on: 17 Aug '11Published in: International Journal of Occupational Medicine and Environmental Health



Abstract

The objective of this study was to investigate variants of case-crossover design for assessing correlations between counts of health events over time and exposure to ambient air pollution. For illustrative purposes, daily emergency department (ED) visits for depression among females were considered.Ambient nitrogen dioxide (NO(2)) was used as a principal ambient air pollutant. In addition, sulphur dioxide (SO(2)) and carbon monoxide (CO) were considered. Different configurations of the control periods (every 1, 2, …, 10 days) and different forms (linear, natural splines) of meteorological factors in the applied conditional logistic regression models were used. The sequence of overlapping age intervals was defined ([0, 19], [1, 20], and so on) and each age group was analyzed separately. The results for the defined age sequences allow identifying age ranges in which the effects are strongest.Consequently, for example, different age ranges for patients for which ED visits for depression were correlated with NO(2) and SO(2) were identified. This age-interval difference explains the very often observed phenomenon whereby two air pollutants used in one model may not show correlations with health outcomes. In this situation they affect separate age groups. The results also show dependency on number-defined control periods for the applied case-crossover technique. The opposite statistical conclusions may be generated by using different control schemas.The results support the hypothesis that ED visits for depressive disorder may be correlated with exposure to ambient air pollution.