481 - The Maternal Vulnerability Index is associated with infant mortality in the United States

Poster Number: 481
Publication Number: 481.129

Daria C. Murosko, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Joshua Radack, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Brielle Formanowski, Childrens Hospital of Philadelphia, Manville, NJ, United States; Molly Passarella, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Elizabeth G. Salazar, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Diana Montoya-Williams, Children’s Hospital of Philadelphia, Philadelphia, PA, United States; Carolyn M. McGann, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Kathryn Paul, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Michelle-Marie Peña, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Heather Burris, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Sara C. Handley, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Scott A. Lorch, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States

Background: The Maternal Vulnerability Index (MVI) is a novel index that examines 43 indicators along six themes (reproductive healthcare, physical health, mental health and substance abuse, general healthcare, socioeconomic determinants and the physical environment) to quantify susceptibility to poor pregnancy outcomes. While MVI has been associated with maternal mortality risk, the association of MVI with infant mortality (IM, death before 1 year of age) has not been examined.

Objective: To examine the association of MVI with IM.

Design/Methods: This retrospective cohort study analyzed national linked birth-death certificate data from 2012-2014 of singleton infants born at 220/7 - 446/7 weeks (or 400–8000g if gestational age [GA] was unavailable). The primary outcome was IM. Multivariate logistic regression determined the association between a 20-point change in MVI and risk of IM clustered at the county level, using fixed effects by state, after adjusting for maternal demographics and health indicators, infant GA and rurality. We also evaluated the association of IM with each of the 6 sub-themes. To assess concordance between MVI and IM at the county level, we calculated the kappa statistic between MVI and IM dichotomized at the national median values.

Results: Of the 11,456,232 infants included in the cohort, 59,056 (0.52%) died before 1 year (Table 1). In the unadjusted model, a 20-point increase in the overall MVI was associated with higher odds of IM (OR 1.18, 95% CI 1.16-1.20); this association persisted in the adjusted model (OR 1.06; 1.04-1.07). IM was associated with all themes in the adjusted models except for the physical environment, and was most strongly associated with physical health, mental health and substance use, and socioeconomic determinants (Table 2). Overall, there was 60.8% agreement between the dichotomized MVI and IM across the US (Figure 1) (k = 0.217); agreement was similar by region and rurality. Conclusion(s): MVI is associated with IM in the United States. After adjusting for perinatal characteristics, this association persisted, indicating that MVI may be a useful proxy measure when seeking to adjust for maternal characteristics in the absence of individual-level maternal data. MVI may also prove useful for capturing typically unmeasured geographic-based factors, particularly aspects of community-level physical health, mental health, substance use and socioeconomics. Further research is needed to better understand the local influences leading to discordance between MVI and mortality at the individual county level.
Table 1. Characteristics of maternal-infant dyads by MVI category**Indicates p < 0.001. The Adequacy of Prenatal Care Utilization Index describes the adequacy of received prenatal care using birth certificate data. Inadequate indicates < 50% of expected visits, intermediate indicates 50-79% of expected visits, adequate indicates 80-109% of expected visits, and adequate+ indicates 110% or more expected visits.
Table 2. Association between MVI^ and infant mortality, overall and stratified by MVI theme**Indicates p < 0.001, *p < 0.05 ^MVI score divided by 20 for interpretability. Adjusted for maternal age, race and ethnicity, insurance, adequacy of prenatal care, nulliparous status, pre-existing hypertension, BMI, smoking, infant sex, gestational age, state, year and urban influence