OBJECTIVE: To assess pregnancies that could have been averted through improved access to contraceptive methods in the 2 years after delivery. METHODS: In this cohort study, we interviewed 403 postpartum women in a hospital in Austin, Texas, who wanted to delay childbearing for at least 2 years. Follow-up interviews were completed at 3, 6, 9, 12, 18, and 24 months after delivery; retention at 24 months was 83%. At each interview, participants reported their pregnancy status and contraceptive method. At the 3- and 6-month interviews, participants were also asked about their preferred contraceptive method 3 months in the future. We identified types of barriers among women unable to access their preferred method and used Cox models to analyze the risk of pregnancy from 6 to 24 months after delivery. RESULTS: Among women interviewed 6 months postpartum (n377), two thirds had experienced a barrier to accessing their preferred method of contraception. By 24 months postpartum, 89 women had reported a pregnancy; 71 were unintended. Between 6 and 24 months postpartum, 77 of 377 women became pregnant (20.4%), with 56 (14.9%) lost to follow-up. Women who encountered a barrier to obtaining their preferred method were more likely to become pregnant less than 24 months after delivery. They had a cumulative risk of pregnancy of 34% (95% confidence interval [CI] 0.25-0.43) as compared with 12% (95% CI 0.05-0.18) for women with no barrier. All but three of the women reporting an unintended pregnancy had earlier expressed interest in using long-acting reversible contraception or a permanent method. CONCLUSION: In this study, most unintended pregnancies less than 24 months after delivery could have been prevented or postponed had women been able to access their desired long-acting and permanent methods.