Variation in use of corticosteroids among infants < or = 1,500 grams across hospitals in three states.

We identified factors associated with no antenatal corticosteroid treatment among 1,369 women who delivered infants < or = 1,500 g and < 34 weeks gestation, 1991-1993. At four hospitals, infants weighing 500-1,500 g were enrolled. Information regarding corticosteroid use, maternal characteristics, and perinatal events were obtained. Factors associated with no corticosteroid treatment were examined in unadjusted and adjusted logistic regression models. Overall, 693 (50.6%) women did not receive corticosteroids. Two hospitals had higher rates of no corticosteroid usage (89% and 75%) as compared with the other two (32% and 50%). Black, Hispanic, and poor women were more likely to receive care at the hospitals where the rates of corticosteroid utilization were lower. Factors in the multivariate model included: < 1 or 1 day of hospitalization, vs. delivery on > or = 2 days of hospitalization (21.4: 14.5, 97.2; 4.7: 3.2, 6.9); gestational age < 26 weeks (2.7: 1.8, 4.1) or > 28 weeks (1.8: 1.3, 2.6) vs. 26-28 weeks; < 12 hours of labor vs. > 12 hours (1.7: 1.2, 2.4); delivering at hospital 2, 3, 4 vs. hospital 1 (1.6: 1.1, 2.5; 24.3: 13.6, 43.4; 10.2: 6.8, 15.3). We conclude that variations in hospital practice limit widespread use of this important antenatal treatment.