Residential mobility during pregnancy and the potential for ambient air pollution exposure misclassification.

Studies of environmental exposures and adverse birth outcomes often rely on maternal address at birth obtained from the birth certificate to classify exposure. Although the gestational age of interest is often early pregnancy, maternal addresses are not available for women who move during pregnancy when using maternal addresses abstracted from birth certificates. The aim of this study was to explore the extent of ambient air pollutant exposure misclassification due to maternal residential mobility during pregnancy among the subgroup of a New York birth cohort. The authors obtained the maternal addresses at birth from the New York Birth Certificate, and the maternal addresses by gestational age from the National Birth Defect Prevention Study for New York participants for the study period 1997-2002. Among the 1324 mothers, 172 (13.0%) moved once during pregnancy and 46 (3.5%) moved at least twice. When accounting for multiple addresses among some individuals, of the 218 mothers who moved, 38 (2.9%) moved in the 3rd to 8th weeks after conception (critical period, not exclusive from the 1st trimester), 80 (6.0%) moved in the 1st trimester, 112 (8.5%) in the 2nd trimester, and 51 (3.9%) in the 3rd trimester. Air monitoring data from the New York Department of Environmental Conservation were used as surrogates to compute the ambient ozone and PM(10) exposures for mothers with complete residential data. This study estimates exposure using maternal address at birth obtained from birth certificates, compared to exposure estimates when using maternal addresses by gestational age obtained from maternal interview, the gold standard. Average exposures during pregnancy were similar when using interview based versus birth certificate addresses (0.035 vs. 0.035 ppm for ozone, and 20.11 vs. 20.09 microg/m(3) for PM(10), respectively). Kappa statistics and percent agreement were calculated to measure the degree of agreement for dichotomous exposure measurements (=median) and weighted kappa for quartile exposure measurements by gestational age. All the statistics indicated a high agreement between the two measurements. For mothers who moved, the majority maintained their address in the same exposure region. Given the low mobility during pregnancy and the short distance moved, the exposure assignment did not change substantially when using the more accurate interview based addresses in this study. However, the level of observed agreement may decrease for studies that require smaller geographic zones for exposure assignments or with more mobile study populations.