Because maternal risk remained statistically similar between 1995 and 2004, these findings raise the question is likely multifaceted. Part of the answer may be that the administrative data used by Bateman et al. did not include maternal characteristics, such as obesity, that affect rates of uterine atony or take into account that coding practices may have changed to better identify maternal hemorrhage. In addition, these data did not include information on the changes in clinical practices that are known to expose more woman to higher rates of uterine atony , such as increased rates of oxytocin induction (indicated and not medically indicated induction).