Multilevel social factors and NICU quality of care in California

Persistent racial/ethnic inequities in adverse birth outcomes are well-documented, yet progress in improving equity in care and outcomes remains limited. These inequities often worsen with the severity of the outcome; for instance, Black infants face a doubled risk of low birth weight and a tripled risk of very low birth weight (VLBW) compared to White infants. VLBW is a major contributor to neonatal mortality. While research has traditionally focused on social determinants' impact on health outcomes, recent studies emphasize that neonatal intensive care unit (NICU) quality of care delivery can either exacerbate or mitigate disparities by identifying higher-risk individuals. Despite overall improvements in outcomes and equity, inequities in care delivery persist. Vulnerable populations are often concentrated in lower-quality NICUs, and within NICUs, they tend to receive lower quality care, especially less family-centered care. Several initiatives utilize quality improvement (QI) strategies to address equity in care, such as a statewide improvement collaborative focusing on disparities in mother's milk provision and the Vermont Oxford Network's (VON) inclusion of equity-relevant aims in their collaboratives. The California Perinatal Quality Care Collaborative (CPQCC) has launched a health equity task force and introduced a health equity dashboard to assess areas of concern within individual NICUs. This dashboard initially focuses on measures previously linked to disparate care; however, it lacked data on social factors like maternal social status, neighborhood conditions, and NICU catchment area characteristics. This study aims to address this gap by incorporating multilevel social factors data to provide contextual information on social determinants of health challenges faced by families before, during, and after their infant's NICU stay. Previous research has linked neighborhood factors like racial residential segregation, income inequality, greenspace, socioeconomic status (SES), and the built environment to birth outcomes. However, the mechanisms by which these factors influence NICU quality of care are not well understood. Social and built environments can affect pregnancy outcomes and mortality through various pathways, including access to high-quality NICUs, social resources affecting mothers' ability to advocate for themselves and their child, and other barriers like language and transportation. For example, neighborhood deprivation can impact breastfeeding rates due to knowledge levels, social support, work demands, or stress. The CPQCC plans to enhance the health equity dashboard with NICU catchment area data (census tract-level data) to provide providers with granular information about their populations' neighborhood challenges and their impact on care.

The introduction section extensively reviews the existing literature on racial and ethnic inequities in birth outcomes, highlighting the disproportionate impact on Black infants. It cites studies demonstrating persistent disparities in NICU quality of care and the concentration of vulnerable populations in lower-quality facilities. The role of social determinants of health is discussed, referencing research linking neighborhood factors (racial segregation, income inequality, access to greenspace, SES, and built environment characteristics) to birth outcomes. The review also acknowledges the lack of understanding regarding the pathways through which these social and environmental factors affect NICU quality of care, mentioning the potential influence on access to high-quality care, maternal advocacy, and other barriers such as language or transportation issues. Finally, it points to existing quality improvement (QI) initiatives aimed at addressing equity in care, including the work of the Vermont Oxford Network (VON) and the California Perinatal Quality Care Collaborative (CPQCC), setting the stage for the current study's focus on integrating social and environmental data into the CPQCC's health equity dashboard.

This study used data from the California Perinatal Quality Care Collaborative (CPQCC) data registry, encompassing VLBW infants born between 2008 and 2011. The initial sample included 19,194 infants, but after applying exclusion criteria (gestational age, birth weight, mortality, congenital abnormalities, and missing maternal race/ethnicity data), the final cohort consisted of 15,901 unique infants from 119 NICUs. The study calculated maternal racial/ethnic and educational composition measures for each NICU. It determined the percentage of infants admitted to their nearest NICU using proximity analyses in ArcMap. Maternal addresses were geocoded using SAS 9.4 to obtain latitude and longitude coordinates and assign 2010 census tract identifiers. Neighborhood-level data on social and built environment attributes were appended from the California Neighborhoods Data System and the American Community Survey. Several factors were considered, including racial/ethnic composition, SES (using a composite measure based on education, housing cost, employment, occupation, income, and poverty), population density, commute patterns, household crowding, street connectivity (walkability using the gamma index), parks, business activity, retail food environment index, restaurant environment index, and traffic density. NICU catchment areas were defined by ranking census tracts based on the number of mothers/infants residing in each tract, selecting tracts encompassing at least 80% of the infants served by each NICU. Census tracts without infants but within the catchment area were included to ensure contiguous areas. NICU quality of care was measured using the Baby-MONITOR score, a composite indicator of nine risk-adjusted measures: antenatal steroid administration, hypothermia on admission, pneumothorax, infection, oxygen requirement at 36 weeks, retinopathy screening, discharge on human milk, mortality, and growth velocity. Statistical analyses involved comparing NICU-level social and built environment factors across tertiles of Baby-MONITOR scores using Pearson's chi-squared tests and analysis of variance.

The study found a strong association between multilevel social factors and NICU quality of care. NICUs with a higher proportion of minority racial/ethnic patients and lower socioeconomic status (SES) patients had significantly lower Baby-MONITOR scores, indicating lower quality of care. Similarly, NICUs located in catchment areas with lower SES, higher minority resident composition, and greater household crowding also exhibited lower quality scores. These findings highlight the significant impact of both the patient population's characteristics and the surrounding community's social and built environment on the quality of care delivered in the NICU. The results underscore the need to consider these multilevel factors when assessing and improving NICU quality and addressing health inequities.

The findings of this study strongly support the hypothesis that social and built environment conditions within a NICU's catchment area are significantly associated with the quality of care provided, as measured by the Baby-MONITOR score. The observed disparities in quality of care between NICUs serving different patient populations and those located in diverse socioeconomic contexts underscore the importance of incorporating multilevel social factors into quality improvement initiatives. The results emphasize the need to move beyond solely focusing on clinical factors and to address the broader social determinants of health that influence both the health of infants and the resources available to their families and the NICUs serving them. This research reinforces the necessity of integrating social and environmental data into health equity dashboards to provide providers with actionable information regarding their patients' resource needs. The significant correlations between neighborhood-level factors (SES, minority composition, crowding) and NICU quality scores suggest that interventions targeting these social determinants may be crucial in improving overall care quality and reducing health inequities. The integration of this data into the CPQCC's dashboard is a significant step towards achieving more equitable outcomes in neonatal care.

This study demonstrates a strong association between multilevel social factors and the quality of neonatal intensive care. NICUs serving higher proportions of minority and low-SES patients, and those located in socioeconomically disadvantaged neighborhoods, showed significantly lower quality scores. Integrating these social determinants into health equity dashboards can empower providers to better understand and address their patients' needs, promoting more equitable care. Future research should investigate specific mechanisms linking these social factors to NICU processes and outcomes and test interventions to improve health equity in neonatal care.

While this study reveals a strong association between multilevel social factors and NICU quality, it is observational and cannot establish causality. The reliance on existing databases may introduce limitations due to data quality and missing values. Additionally, the study focuses on a specific geographic area (California) and a particular time period, potentially limiting generalizability to other settings or time points. Further research is needed to explore potential mediating factors and develop targeted interventions.