ABSTRACT The US is in the midst of a maternal health crisis. The US has worse maternal mortality (MM) than any other high-income country and is the only one for which the rate is increasing. Severe maternal morbidity (SMM), which encompasses conditions that put pregnant women most at risk of dying, doubled in the last two decades. Women of color, especially Black and Native American women, are at 2-3-fold or greater increased risk of these outcomes. This proposal focuses on SMM, as a sentinel outcome leading to MM, yet 100 times more common. Most prior research on SMM has focused on proximal clinical factors (primarily related to co-morbidities and obstetric management); these factors alone are insufficient for explaining SMM or its racial/ethnic disparities. The objectives of this proposal are: 1) develop a causal pathway framework to understand how social determinants and more proximal health-related factors together contribute to SMM risk and disparities, and 2) use this framework to identify actionable strategies to reverse current trends and eliminate inequities. We will create a unique dataset that harmonizes 4 years of data from 6 states on 4.4 million births and 66,000 women with SMM. These states collectively include 1 in 4 US births and sufficiently diverse social environments to disentangle complex multi-level drivers of maternal health inequity. Our focus is on social determinants from 4 specific domains: socioeconomic disadvantage, structural inequality, community resources, and health care access, characterized at the county, neighborhood (census tract or ZIP code), and individual level. Health-related intermediaries include birth hospital quality of care, mode of birth, and maternal morbidities (eg, hypertension, anemia). These domains and intermediaries were selected for their relevance to racial/ethnic disparities in maternal health. In addition, we will evaluate impacts of 3 Quality Improvement (QI) collaboratives on SMM disparities, which were designed to improve specific aspects of hospital quality of care and implemented by state-wide perinatal quality collaboratives (PQCs) in 3 of the 6 states included in this proposal. Our Specific Aims are: 1) Assess the relative and joint contributions of multi-level social determinants to SMM and SMM disparities; 2) Identify potential health-related mechanisms by which multi-level social determinants affect SMM and SMM disparities by conducting causal pathway analyses; and 3) Evaluate the impact of 3 QI collaboratives designed to reduce SMM, on racial/ethnic disparities in SMM (3a), and apply the causal framework developed in Aims 1 and 2 to these state-specific contexts, to identify state-specific strategies for addressing SMM disparities (3b). We currently have very limited understanding of how social determinants contribute to SMM or its disparities. The proposed research will fill this important knowledge gap, which is an essential part of realizing sustainable improvement of maternal health and elimination of racial/ethnic disparities.