The West African country of Mali implemented referral systems to increase spatial access to emergency obstetrical care and lower maternal mortality. We test the hypothesis that spatial access- proxied by travel time during the rainy and dry seasons- is associated with in-hospital maternal mortality. Effect modification by caesarean section is explored. All women treated for emergency obstetrical complications at the referral hospital in Kayes, Mali were considered eligible for study. First, we conducted descriptive analyses of all emergency obstetrical complications treated at the referral hospital between 2005 and 2007. We calculated case fatality rates by obstetric diagnosis and travel time. Key informant interviews provided travel times. Medical registers provided clinical and demographic data. Second, a matched case-control study assessed the independent effect of travel time on maternal mortality. Stratification was used to explore effect modification by caesarean section. Case fatality rates increased with increasing travel time to the hospital. After controlling for age, diagnosis, and date of arrival, a travel time of four or more hours was significantly associated with in-hospital maternal mortality (OR: 3.83; CI: 1.31-11.27). Travel times between 2 and 4 h were associated with increased odds of maternal mortality (OR 1.88), but the relationship was not significant. The effect of travel time on maternal mortality appears to be modified by caesarean section. Poor spatial access contributes to maternal mortality even in women who reach a health facility. Improving spatial access will help women arrive at the hospital in time to be treated effectively.