Drs Reichman and Samueloff raise 4 issues: (1) our inclusion criteria, (2) the use of painful uterine contractions and Bishop score as possible predictors of true vs false labor, (3) sample size concerns, and (4) selection bias. Below are our responses.

First, our inclusion criteria (cervix <4 cm dilated and <80% effaced) agree with the traditionally accepted definitions of labor, allowing for inclusion of patients who could have been in either the latent phase of true labor or in false labor. Similar inclusion criteria have been used in previous studies, facilitating comparison of results.

Second, Drs Reichman and Samueloff opined that our inclusion criteria were “nonspecific” and that we should have used painful uterine contractions and Bishop score because these could theoretically predict better true vs false labor. We believe that the classification of uterine contractions as painful or painless is nonspecific since Drs Reichman and Samueloff do not provide any measure or method how to reliably quantitative the pain. Bishop score is used to determine the likelihood of vaginal delivery in patients before labor induction and it has not been shown to be clinically helpful in differentiating the latent phase of early true labor from false labor. We would welcome a future study comparing the accuracy of cervical length vs painful contractions and Bishop score for predicting true or false labor as such a study would provide an empirical basis for evaluating the conjecture of Reichman and Samueloff that the latter combination of variables would be better predictors.

Third, our study was powered to address the primary objective, namely, to determine if cervical length can differentiate true from false labor irrespective of parity. The stratified analysis based on parity was a secondary objective. We disagree that combining the groups produces misleading results, but agree that it would be useful to conduct a larger study to examine the effect of parity.

Fourth, the reason for enrolling nonconsecutive patients was addressed in our article (time constraints due to a very busy labor and delivery practice). Patients were enrolled based on eligibility and availability of an investigator to enroll the patient. While this approach did not guarantee a representative sample–since we did not enroll a consecutive series of patients or use random sampling– systematic bias seems unlikely. As with any nonrandom sample, the possibility of nonrepresentativeness cannot be ruled out.