Preterm birth prevention: evaluation of a prospective controlled randomized trial.

Over a 3-year period 5457 indigent patients were scored for risk of preterm birth and 4595 women were delivered at greater than or equal to 20 weeks' gestation. Patients at high risk (18.1%) were randomized into control and intervention groups. The latter group received weekly cervical examinations and instruction regarding subtle symptoms and signs of preterm labor. Medical providers received similar instruction. There was no difference in preterm births between control and intervention groups (20.8% vs. 22.1%). Medical providers, convinced of preterm birth prevention during year 1 of the study, defeated the study design by giving preterm birth precautions to all patients. In turn, preterm births decreased from 13.7% (year 1) to 9.3% (year 2, p less than 0.001) and remained stable in year 3 (8.7%). Preterm births during year 1 and the 8 months preceding year 1 were not different. Significant differences in preterm births between private and indigent study patients during these two periods (p less than 0.001) disappeared during years 2 and 3 of the study.