Objectives: This report measured the effect of births at 22 weeks' gestation or earlier on infant mortality in Philadelphia, Pa.
Methods: The proportion of live-born deliveries at 22 weeks or earlier was calculated. Overall and race-specific infant mortality was calculated after excluding live-born deliveries at 22 weeks' gestation or earlier.
Results: Of all deliveries, 1.5% were at 22 weeks or earlier. Of these, 68% were stillborn and 32% were live-born. Large hospital-to-hospital variation in the proportion of live-born deliveries at 22 weeks' gestation or earlier was noted. When nonviable births were excluded, overall infant mortality decreased 40%.
Conclusions: The development of a standardized birth certificate policy is needed and will facilitate comparisons of infant mortality across spatial boundaries and racial/ethnic groups.