This Indonesian study evaluates associations between near-miss status/death with maternal demographic, health care characteristics, and obstetrical complications, comparing results using retrospective and prospective data. The main outcome measures were obstetric conditions and socio-economic factors to predict near-miss/death. We abstracted all obstetric admissions (1,358 retrospective and 1,240 prospective) from two district hospitals in East Java, Indonesia between 4/1/2009 and 5/15/2010. Prospective data added socio-economic status, access to care and referral patterns. Reduced logistic models were constructed, and multivariate analyses used to assess association of risk variables to outcome. Using multivariate analysis, variables associated with risk of near-miss/death include postpartum hemorrhage (retrospective AOR 5.41, 95 % CI 2.64–11.08; prospective AOR 10.45, 95 % CI 5.59–19.52) and severe preeclampsia/ eclampsia (retrospective AOR 1.94, 95 % CI 1.05–3.57; prospective AOR 3.26, 95 % CI 1.79–5.94). Associations with near-miss/death were seen for antepartum hemorrhage in retrospective data (AOR 9.34, 95 % CI 4.34–20.13), and prospectively for poverty (AOR 2.17, 95 % CI 1.33–3.54) and delivering outside the hospital (AOR 2.04, 95 % CI 1.08–3.82). Postpartum hemorrhage and severe preeclampsia/eclampsia are leading causes of near-miss/death in Indonesia. Poverty and delivery outside the hospital are significant risk factors. Prompt recognition of complications, timely referrals, standardized care protocols, prompt hospital triage, and structured provider education may reduce obstetric mortality and morbidity. Retrospective data were reliable, but prospective data provided valuable information about barriers to care and referral patterns.