OBJECTIVE We sought to evaluate whether group B streptococcus (GBS) detection is altered by the digital cervical examination. METHODS A total of 302 women undergoing the clinical GBS culture had a digital cervical examination and a repeated GBS culture. Statistical comparison of pre-post culture results were performed with kappa and McNemar tests. RESULTS The clinical prevalence of GBS was 19.5%. Discordant results were seen in 30/302 (9.9%) paired cultures (kappa = 0.68; 95% confidence interval, 0.568-0.783). An initially negative GBS culture result was positive on repeated testing in 13/243 (5.3%) pairs. Initially positive cultures were negative on repeated testing in 17/59 (28.8%) pairs. Patients with discordant results had similar characteristics as the remainder of the study group. Given the observed proportion of discordant results (9.9%), the study had 80% power to detect a 5% difference between discordant pairs. CONCLUSIONS Paired GBS cultures showed a good level of agreement. The 28.8% rate of positive cultures becoming negative is clinically concerning and warrants further study.