“…This proxy variable includes a substantial and increasing fraction of late-pregnancy terminations for prenatally diagnosed congenital anomalies. 5,26,27,36,37 Data were stratified according to geographic area of maternal residence (14 strata: 5 regions in Ontario; 2 in Alberta; 2 in British Columbia; 1 for Saskatchewan, New Brunswick, and Newfoundland each; 1 for Nova Scotia and Prince Edward Island combined; and 1 for the Northwest Territories, Yukon, and Nunavut combined) and year (9 strata from 1990–1998, 13 strata from 1999–2011), resulting in 126 strata for the prefortification period and 182 strata for the post–food fortification period. Goodness of fit of the Poisson regression model was assessed with deviance statistics and the Pearson χ 2 test, and variance estimates were corrected for overdispersion through appropriate scaling.…”