A nonconcurrent prospective study of 273,454 sigleton deliveries in South Carolina from 1989-1993 was conducted to determine the relationship between fetal mortality and rural status of maternal residence. Four geographic areas were identified based on the population of the largest city in the county(rural [R], moderately rural [MR], moderately urban [MU], and urban [U]). The fetal mortality rate for the R, MR, MU, and U groups were 13.6, 12.1, 10.3, and 8.7, respectively. There were no geographic differences in gravidity. With increasing rurality, there was a linear increase in preterm deliveries, low birth weight, Black ethnicity, single marital status, low education, and teenage pregnancies. Late entry into prenatal care and delivery in large volume hospitals demonstrated `U-shaped' distributions. Controlling for population socio-demographic differences with logistic regression, rural groups remained at increased risk for fetal death (R:OR 1.4; MR:OR 1.3; MU:OR 1.2; U:OR 1.0 ref). The unadjusted fetal mortality rate was increased for deliveries in large volume hospitals in all geographic areas. This was directly related to a four fold increase in the percent of very low birth weight deliveries in these larger hospitals since the VLBW specific mortality was lower for larger hospitals (see figure). Hospital size, location of VLBW deliveries, and rurality appeared to be the most significant factors in differences in fetal mortality. Controlling for hospital size of delivery, there were no statistical differences in fetal mortality by rural status. The adjusted odds ratios were: R:OR 1.1 (0.9-1.3); MR: 0.8 (0.7-1.0); and MU: 0.9 (0.8-1.0). These data suggest that residential differences in birth weight and size of hospital of delivery are significantly associated with geographic differences in fetal mortality.

figure 1

Figure 1