Racial/ethnic disparities in the quality of care provided in neonatal intensive care units (NICUs) to very low- birth-weight (VLBW; 1500g) result directly in avoidable death and morbidity. National health policy recommendations underscore that a comprehensive understanding of the multi-level factors that influence NICU quality, including mother and infant clinical, hospital, sociodemographic, and neighborhood factors, is needed to re- duce disparities. To date, research examining disparities of NICU quality suffers from considerable limitations. First, sociodemographic and neighborhood factors are rarely available for studies of NICU quality so their in- dependent and joint contributions are not well understood. Second, a lack of multi-level adjustment under- mines the validity of comparative assessments of quality by penalizing NICUs that care for vulnerable populations. Third, studies of neighborhood factors have narrowly focused on mortality, yet the majority of infants survive their birth hospitalization. To address these limitations and significantly advance the understanding of disparities in NICU quality, we will undertake a comprehensive appraisal of the multi-level factors associated with NICU quality; promote fairer comparisons by making sociodemographic and neighborhood influences explicit; and assess key areas of quality beyond mortality, including processes and outcomes of care that modulate long-term neu- rodevelopmental outcomes. Innovative methods include leveraging our vanguard efforts in developing and linking unique multi-level data sources to provide comprehensive assessments of NICU quality with unprecedented granularity, assessing NICU quality via key individual and composite measures, and using a mixed-methods approach to pro- vide unique insights into the mechanisms by which sociodemographic and neighborhood factors optimize or degrade NICU quality. This study is designed to achieve 3 specific aims:

1) Test the association of multi-level risk factors with NICU quality,

2) Evaluate the effect of multi-level risk adjustment on comparative ratings of NICU quality; and

3) Understand the drivers of disparities in NICU quality.

An exceptional strength of our proposal is our ability to leverage the unique infrastructure of the California Perinatal Quality Care Collaborative and its national counterpar the Vermont Oxford Network. Building on our pri- or work in linking complimentary high-quality clinical, administrative, and survey-based data sets containing multi-level factors, we will study large population-scale sample of >30,000 VLBW infants in >130 NICUs in California between 2008-16. Results of this proposal are expected to have an immediate positive impact be- because they are designed to identify pragmatic areas where NICUs and communities can improve.