In a retrospective case control study with historical controls, the influence of the introduction of vaginal sonographic cervical length measurement on the utilization of hospital services and pregnancy outcome of gravid women at risk of preterm delivery was examined. Prior to the introduction of vaginal ultrasonography to measure cervical length, we registered 76 hospitalizations totalling 1827 hospitalization days due to premature cervical ripening and/or premature labor in 1991 and 1992. In 1994 and 1995, after vaginal ultrasonography was introduced and intravenous tocolysis was limited to cases with cervical shortening to less than 3 cm, there were 64 admissions resulting in just 869 hospitalization days. Long-term hospitalizations (over 10 days) were reduced from 55 to 25 cases (p < 0.0001) and the median length of hospital stay decreased from 18 to 8 days (p < 0.0001). The number of preterm births (< or = 37 weeks) remained stable: 12 cases in 1991-1992 and 13 cases in 1994-1995. In conclusion, vaginal sonographic measurement of cervical length provides an objective criterion for cervical shortening with premature labor. The method could therefore be a suitable means of reducing unnecessary therapeutic interventions in gravid women with premature contractions and/or cervical dilatation. A prospective randomized trial to confirm these findings is suggested.