Abstract Objective To evaluate the impact of ultrasound cervical length measurement on duration of hospital stay in patients admitted for threatened preterm labor. Study design This was a prospective, comparative study in 294 patients with threatened preterm labor in three hospitalization units (Units A, B and C). In the first phase of the study (observational), cervical length was measured by transvaginal ultrasound, but managing physicians were blinded to the results. In the second phase (interventional), physicians from Unit A remained blinded to cervical length information, but Units B and C incorporated these data into their clinical management protocols. Early discharge was contemplated if the cervix measured 25 mm or more on admission (Unit B) or no changes were observed over 48 h (Unit C). Duration of hospital stay and delivery rates within 7 days and before 37 weeks' gestation were recorded. Results Hospital stay was significantly reduced in Units B and C in the interventional phase, while no changes were observed in Unit A. Delivery rates within 7 days and before 37 weeks' gestation were similar in the three units during the two stages of the study. Conclusion Routine use of ultrasound cervical length assessment in patients admitted with threatened preterm labor may reduce the duration of hospital stay without increasing the rate of preterm births. These data should be confirmed by means of an appropriately designed randomized clinical trial. Copyright © 2004 ISUOG. Published by John Wiley & Sons, Ltd.