Measurement of uterine contraction frequency has been employed as a screening test to identify women with increased risk of preterm birth, and as an aid in the early diagnosis of preterm labor. The National Institute of Child Health and Human Development Maternal-Fetal Medicine Units (NICHD MFMU) Network performed a prospective, blinded observational study of uterine contraction frequency to detect and predict preterm labor and birth, respectively. The goal of the study was to assess the sensitivity, specificity, and positive and negative predictive value of various measures of uterine contraction frequency. Data collected from 306 women revealed that contraction frequency was significantly greater in women who would ultimately deliver before rather than after 35 weeks' gestation. However, both sensitivity and positive predictive value of any measure of contraction frequency to predict preterm birth were poor. Contraction frequency did not increase significantly within 1 or 2 weeks of an episode of preterm labor. These results serve to explain the absence of an association between contraction-based surveillance and preterm birth in randomized trials conducted in women at risk of preterm birth.