The 'continuum of care' is proposed as a key framework for the delivery of maternal, neonatal and child health services. This study examined the extent of dropout as well as factors associated with retention across the MNCH continuum from antenatal care (ANC), through skilled birth attendance (SBA), to postnatal care (PNC). We analyzed data from 1931 women who delivered in the preceding 2-14 months, from a twostage cluster sampling household survey in four districts of Tanzania's Morogoro region. The survey was conducted in 2011 as a part of a baseline for an independent evaluation of a maternal health program. Using the Anderson model of health care seeking, we fitted logistic models for three transition stages in the continuum. Only 10% of women received the 'recommended' care package (4+ANC visits, SBA, and 1+PNC visit), while 1% reported not having care at any stage. Receipt of four ANC visits was positively associated with women being older in age (age 20-34 years - OR: 1.77, 95%CI: 1.22-2.56; age 35-49 years - 2.03, 1.29-3.2), and knowledge of danger signs (1.75, 1.39 -2.1). A pro-rich bias was observed in facility-based deliveries (proxy for SBA), with women from the fourth (1.66, 1.12-2.47) and highest quintiles of household wealth (3.4, 2.04-5.66) and the top tertile of communities by wealth (2.9, 1.14-7.4). Higher rates of facility deliveries were also reported with antenatal complications (1.37, 1.05-1.79), and 4+ANC visits (1.55, 1.14-2.09). Returning for PNC was highest among the wealthiest communities (2.25, 1.21-4.44); catchment areas of a new PNC program (1.89, 1.03- 3.45); knowledge of danger signs (1.78, 1.13-2.83); community health worker counselling (4.22, 1.97-9.05); complicated delivery (3.25, 1.84-5.73); and previous health provider counselling on family planning (2.39, 1.71-3.35). Dropout from maternal care continuum is high, especially for the poorest, in rural Tanzania. Interactions with formal health system and perceived need for future services appear to be important factors for retention.
