A retrospective cohort study was conducted to compare pregnancy outcomes in farming households that used pesticides conventionally with those that practiced integrated pest management (IPM) in Nueva Ecija, Philippines, in the period 1998-1999. Conventional pesticide users (CPUs) were defined as pesticide appliers who used pesticides routinely and regularly, whereas users of IPM were those who used pesticides as necessary, and on economically injured crop areas only. The data sets were subjected to the chi-square test of association, Fisher's exact probability test, and logistic regression analysis. At a significance level at 0.05, spontaneous abortion occurred significantly more often among the 345 CPU households than among the 331 IPM households (adjusted risk ratio 6.17). Likewise, birth defects were significantly more common in the CPU group (adjusted risk ratio 4.56). Thus, people of reproductive age who plan to have children should avoid any use of pesticides.