Objectives Timely access to health care is critical in obstetrics. Yet obtaining reliable estimates of travel times to hospital for childbirth poses methodological challenges. We compared two measures of travel time, self-reported and calculated, to assess concordance and to identify determinants of long travel time to hospital for childbirth. Methods Data came from the 2010 French National Perinatal Survey, a national representative sample of births (N = 14 681). We compared both travel time measures by maternal, maternity unit and geographic characteristics in rural, peri-urban and urban areas. Logistic regression models were used to study factors associated with reported and calculated times ≥30 min. Cohen's kappa coefficients were also calculated to estimate the agreement between reported and calculated times according to women's characteristics. Results In urban areas, the proportion of women with travel times ≥30 min was higher when reported rather than calculated times were used (11.0 vs. 3.6%). Longer reported times were associated with non-French nationality [adjusted odds ratio (aOR) 1.3 (95% CI 1.0-1.7)] and inadequate prenatal care [aOR 1.5 (95% CI 1.2-2.0)], but not for calculated times. Concordance between the two measures was higher in peri-urban and rural areas (52.4 vs. 52.3% for rural areas). Delivery in a specialised level 2 or 3 maternity unit was a principal determinant of long reported and measured times in peri-urban and rural areas. Conclusions for Practice The level of agreement between reported and calculated times varies according to geographic context. Poor measurement of travel time in urban areas may mask problems in accessibility.