To characterize the continuum between normal and preterm labor to prevent spontaneous preterm birth and low birth weight. Between August 1, 1992, and November 30, 1993, obstetric patients from a large managed care medical group were assigned in rotation to five private practice offices for pregnancy care. In the study office (374 births), a systematic approach of visit-by-visit screening, patient education, selective office evaluation of cervical change by transvaginal ultrasound examination, outpatient modification of activity, and graded oral or subcutaneous tocolysis was used; inpatient tocolysis was used only when outpatient management failed. In the comparison offices (1391 births), transvaginal ultrasound was not used and conventional methods were used at the discretion of the attending physicians. Birth weights below 1500 g (P = .008; odds ratio [OR] 0.08, 95% confidence interval [CI] 0.05–1.32), 2000 g (P = .21; OR 0.21, 95% CI 0.05–0.88), and 2500 g (P = .008; OR 0.44, 95% CI 0.23–0.83) occurred significantly less often in the study group than in the comparison group. The difference in spontaneous preterm births under 2500 g was also significant (P < .001; OR 0.08, 95% CI 0.01–0.58). By emphasizing transvaginal ultrasound and graded outpatient tocolysis, the diagnosis and management of preterm prelabor was associated with a reduction in the rate of spontaneous preterm birth and low birth weight infants.