Our purpose was to compare physician opinions about the appropriate use of corticosteroids with physician perceptions of preterm infant outcome. A total of 409 obstetricians and family physicians who provide maternity care in Alabama were surveyed to determine whether (and how) their perceptions of preterm infant outcome influenced their decision to use antenatal corticosteroids. Results were compared with those of a similar survey of Alabama physicians conducted in 1979. A total of 85% of physicians in 1992 versus 61% in 1979 reported that situations existed in which they would prescribe antenatal corticosteroids (p < 0.001). In 1992 physicians who underestimated preterm infant survival began corticosteroid use at later gestational ages than those physicians who did not underestimate survival (p < 0.02). In addition, 54% of physicians who underestimated preterm infant survival reported giving corticosteroids between 23 and 28 weeks' gestational age versus 73% of physicians with more accurate perceptions (p < 0.03). We conclude that in 1992, compared within 1979, more Alabama physicians reported antenatal corticosteroid use. Use of this treatment was influenced by each physician's perceptions of the probable outcome of the preterm infant.