Objective To describe the results of improving availability of, and access to emergency obstetric care (EmOC) services in underserved rural Tanzania. Design Prospective cohort study Settings Rural Tanzania Methods Forty two associate clinicians from five health centres were trained in teams for three months in emergency obstetric care and anaesthesia. Two health centres were unexposed to the intervention and served as controls. Following training, virtual teleconsultation, quarterly on-site supportive supervision and continuous mentorship were implemented to reinforce skills and knowledge. Main outcome measures Proportion of all births in emergency obstetric care facilities, met need for EmOC and case fatality rate. Results The met need for EmOC increased significantly from 45% (459/1,025) at baseline (July 2014 – June 2016) to 119% (2,010/1,691) during the intervention period (Jul 2016 – June 2019). The met need for EmOC in the control group also increased from 53% (95% CI 49%-58%) to 77% (95% CI 74%-80%). Forty maternal deaths occurred during the baseline and intervention periods in the control and intervention health centres. The direct obstetric case fatality rate decreased slightly from 1.5% (95% CI 0.6%–3.1%) to 1.1% (95% CI 0.7%–1.6%) in the intervention group and from 3.3% (95% CI 1.2%–7.0%) to 0.8% (95% CI 0.2%–1.7%) in the control group. Conclusions When EmOC services are made available the proportion of obstetric complications treated in the facilities increases. However, the effort to scale up EmOC services in underserved rural areas should be accompanied by strategies to reinforce skills and the referral system.