Design considerations in pregnancy outcome studies of occupational populations.

Increased attention has been focused on the relationship between adverse pregnancy outcomes and occupational exposures to parents of both sexes. However, the characteristics and dynamics of working populations impose limitations on potential study populations. These limitations result from the combined effects of workforce size, exposure, age distribution, and marital status, which limit the number of pregnancies available for study. The smaller populations available typically result in retrospective studies covering extended time periods. Potential data sources for these studies include interview data, medical records, vital statistics data, and insurance records. All sources may have biased ascertainment of certain pregnancy outcomes such as early fetal loss due to errors in recall in interview data, legal requirements for recording vital statistics data, and differences in medical care patterns.