Studies of cervical sonography, fetal fibronectin, and uterine contraction monitoring during pregnancy have improved our understanding of how preterm labor occurs, but their use in practice remains uncertain. This article reviews the use of these tests to improve the accuracy of diagnosis of preterm labor and to estimate the likelihood of preterm birth in women with historical risk factors. Because the clinical criteria for a diagnosis of preterm labor are inaccurate until labor is well established, over-diagnosis is common. A cervical length measurement of more than 30 mm or a negative fibronectin obtained from a patient with possible preterm labor can avoid over-diagnosis and unnecessary treatment. There is no role for routine use of either cervical sonography or fibronectin to screen pregnant women for preterm birth risk, but women thought to be at increased risk may be reassured by negative test results. Uterine contraction monitoring has low sensitivity for detecting women at risk of preterm birth. Current use of cervical length and fetal fibronectin in pregnancy is limited to situations where a negative result can avoid unnecessary interventions.