OBJECTIVE: To describe caesarean section rates and neonatal mortality to assess change in access to life-saving interventions in a rural low-resource setting between 2007 and 2013. DESIGN: Population-based cross-sectional study. SETTING: Southern Tanzania. POPULATION: A total of 34 063 women from 384 549 households who gave birth in the previous year. METHODS: Using data collected in two geo-referenced household surveys conducted in 2007 and 2013 in the context of two cluster-randomized controlled trials, we describe trends in caesarean section and neonatal mortality in obstetric risk groups inspired by the 10-group Robson classification. MAIN OUTCOME MEASURES: Rates of self-reported birth by caesarean section and neonatal mortality. RESULTS: Population-based caesarean section rates increased from 4.0% in 2007 to 6.4% in 2013. In 2013, the lowest caesarean section rate was found in multipara whose labour was not induced or augmented [4.4%, 95% confidence interval (CI) 3.9-4.9], a group that showed a rate increase of over 50% from 2007 [adjusted prevalence ratio 1.57 (95% CI 1.34-1.82)]. Nullipara whose labour was not induced or augmented had rates of 6.2% in 2007 and 8.5% in 2013. Caesarean rates in multiple pregnancies were low at 8.1% (95% CI 5.6-10.5) in 2007, and 14.6% (95% CI 9.4-19.8) in 2013. Overall neonatal mortality was high: 3.5% in 2007 and 3.2% in 2013, with rates being lowest in multiparous women whose labour was not induced or augmented: 2.4% (95% CI 2.2-2.7) and 1.7% (95% CI 1.4-2.0) in 2007 and 2013, respectively. CONCLUSION: Although use of caesarean section remains insufficient, and higher rates do not necessarily imply better quality of care, our analysis highlights improvements in reaching women with caesarean section. Rates in multiple birth remained low compared with high-income settings. TWEETABLE ABSTRACT: In Southern Tanzania caesarean section rates increased over time, but the rate in high-risk births remained alarmingly low.