New Jersey maternal mortality dashboard: an interactive social-determinants-of-health tool

Introduction

Maternal mortality is a key measurement that reflects the health status of a population and the quality of care that population receives [1, 2]. Women in the United States are more likely than those in other developed countries to die from complications related to pregnancy or childbirth [3, 4]. We recognize that more inclusive terms for individuals who give birth are preferable (e.g., birthing people). However, available data and statistical reports use the term “women.” Therefore, the term “women” is used herein.

Maternal mortality is surveilled in the United States using the National Center for Health Statistics (NCHS) National Vital Statistics System (NVSS) and the Pregnancy Mortality Surveillance System (PMSS), which are supported by the Centers for Disease Control and Prevention (CDC), as well as state-based Maternal Mortality Review Committees (MMRCs) [5], [6], [7], [8]. A checkbox for pregnancy-related deaths on United States Standard Certificates of Death was initiated in 2003 and took until 2018 to be implemented in all 50 states, the District of Columbia, and United States territories [9, 10]. In 2018, the NCHS released its first annual report of national estimates of maternal mortality since 2007 [10], utilizing data from the death certificate check box. According to the 2019 report [5], the overall maternal mortality rate (deaths per 100,000 live births) was significantly higher in 2019 (20.1; 754 deaths) than in 2018 (17.4; 658 deaths). These findings are inconsistent with prior reports that maternal mortality rates did not increase significantly between 2002 and 2018 (except 2017) [9], [10], [11], [12]. Findings relative to race/ethnicity, though, were consistent with other reports [6, 10, 13, 14]. The maternal mortality rates in 2018 and 2019 for non-Hispanic Black women (37.3 and 44.0, respectively) were significantly higher — 2.5- to 3.5-times — than for non-Hispanic White (14.9 and 17.9) or Hispanic (11.8 and 12.6) women [5]. The maternal mortality rate in 2019 also significantly increased with age, from 12.6 for women <25 years old, to 19.9 for women aged 25 to 39, to 75.5 for women ≥40 years old. However, only the rate in the 25–39 age group significantly increased from 2018 (16.6) to 2019 (19.9) [5].

Interpreting United States maternal mortality data is challenging, however, because of varied sources, inconsistent reporting, and poor quality of data, as well as diverse terms and definitions (Table 1) [5], [6], [7], [8]. Moreover, MMRCs have access to additional information on maternal deaths (e.g., medical and social records), which allows a more in-depth evaluation of factors leading to the deaths than that available from vital registration information alone [6, 7].

Maternal mortality rates vary greatly across states, adding to the complexity of interpreting maternal mortality data. According to the most recent state-level NCHS NVSS data (2018), maternal deaths ranged from 9.7 per 100,000 live births in Illinois to 45.9 in Arkansas [17], suggesting that individual states face unique challenges and, therefore, may need tailored solutions. Healthcare professionals, non-profit organizations, public health officials, and policymakers are working to understand the drivers of maternal mortality [18]. At the federal level, the CDC has the Pregnancy Risk Assessment Monitoring System (PRAMS) [19], the Enhancing Reviews and Surveillance to Eliminate Maternal Mortality (ERASE MM) program [20], and the CDC Levels of Care Assessment Tool^SM (CDC LOCATe^SM) [21], and the National Institutes of Health has the Implementing a Maternal health and PRegnancy Outcomes Vision for Everyone (IMPROVE) initiative [22].

An interactive data visualization tool for maternal health-related information in New Jersey, which has one of the highest pregnancy-related maternal death rates in the United States (26.7 per 100,000 live births in 2018) [17] was developed, and we explored how the tool could be used. Here, we report the process for development of the New Jersey Maternal Mortality Dashboard (NJMMD) and illustrate the how the tool enables users to simply and graphically depict cross-sectional, high-level perspectives on potential correlations between maternal mortality with demographic, social, and public health factors at a state and county level. We are sharing the tool broadly so other states or regions have the opportunity to use it for their own geographies.

Methods

Characteristics of the NJMMD tool

The NJMMD includes multiple measures, allowing the user to have a broader view of maternal mortality based on different data points. Maternal mortality, measured as MMRates or MMRatios, are available by year (2005–2017) for all women, by age (in 5-year blocks from 15 to 49 years of age; Figure 1), and by race/ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, Hispanic, or other). MMRates or MMRatios, along with the contextual variables, can be visualized on a map that includes all New Jersey counties. Because of its interactive nature, trends and patterns may be observed based on these data points. Questions to be addressed include: Is there a match or disconnect when looking at the data? Are there sufficient and appropriate types of care available based on the data?

Figure 1:

Examples of the ability to evaluate data based on age of mother (top) or race/ethnicity of mother (bottom) using the New Jersey Maternal Mortality Dashboard tool.

Results

Use cases of the NJMMD tool

Overall MMRatios (deaths per 100,000 live births) in New Jersey varied over time but trended in an upward manner (mean [standard deviation (SD)]: 36.7 [±10.4]; Figure 2). MMRatios by counties also varied from year to year. However, in Essex county, which has one of the highest number of delivery hospitals and obstetrician/gynecologists per capita in the state (Figure 3), MMRatios generally were high over time (mean [SD], range: 64.0 [±28.2], 30–111 annually), as were poverty rates (mean [SD], range: 16.8% [±0.63%], 15.7–17.7% annually). These types of data, combined with available data from hospitals and birthing sites in the county, and prenatal care sites (e.g., Federally Qualified Health Centers), can inform planning for not only provision of care (e.g., identify gaps in types, locations, and inequities of resources, as well as communication about resources; identify gaps in types, races, and/or languages of providers and care teams; and clarify barriers to care) but also levels of care (level I [basic care], level II [specialty care], level III [subspecialty care], and level IV [regional perinatal health care centers]) needed.

Figure 2:

Maternal mortality ratio (number of maternal deaths per 100,000 live births) over time (graph), and by county (map) in 2017 in New Jersey. Chart (left) was creating using data from the New Jersey Maternal Mortality Dashboard website, and the graphic containing the map (right) is a screen shot from the New Jersey Maternal Mortality Dashboard website.

Figure 3:

Maternal mortality ratio (number of maternal deaths per 100,000 live births; blue bars on graph) and poverty rates (orange on graph) over time, and number of delivery hospitals, and obstetrician/gynecologists and practitioner midwives per capita (maps) in 2017, in Essex county, New Jersey. Chart (left) was creating using data from the New Jersey Maternal Mortality Dashboard website, and the maps (right) are screens shots from the New Jersey Maternal Mortality Dashboard website.

In 2017, the overall MMRatio in New Jersey was 41.7 (Figure 2). MMRatios were highest in Cumberland (167) and Cape May (133) counties. High MMRatios in Cumberland county likely were driven by high MMRatios among non-Hispanic Black women (579) (Figure 4A). Cumberland county also had the highest rates of poverty (18.4% [Figure 4B] vs. 10.0% for New Jersey [23] and 13.4% for the nation [23]) and Medicaid coverage (33.7% [Figure 4C] vs. 17.2% for New Jersey [24] and 19.3% for the nation [25]), and the sixth highest percentage of uninsured women (13.7% [Figure 4D]) after Passaic county (15.9 vs. 7.7% for New Jersey [26] and 8.7% for the nation [26]). High MMRatios in Cape May county likely were driven by high MMRatios among non-Hispanic White women (176) (Figure 5A). Cape May county had the highest difference in NTSV Cesarian rate (12.9% [Figure 5B]) from the New Jersey statewide 2020 goal of 24.7%. Based on these data, we can hypothesize that the high maternal mortality among non-Hispanic Black women in Cumberland county is associated with high poverty and low insurance coverage, while the high maternal mortality among non-Hispanic White women in Cape May county is associated with high NTSV Cesarian rate. In 2017, the MMRatio was zero in eight New Jersey counties (Warren, Hunterdon, Somerset, Mercer, Burlington, Ocean, Atlantic, and Salem). Poverty rates also were generally low in these counties (mean [SD], range: 9.1% [±3.8%], 3.9–14.4% annually). Collectively, these types of data provide an opportunity to identify trends and areas of focus for further analysis of factors impacting outcomes.

Figure 4:

Maternal mortality ratio (number of maternal deaths per 100,000 live births) (A), percent poverty (B), percent covered by Medicaid (C), and percent uninsured (D) in 2017 in Cumberland county, New Jersey. Maps are screen shots from the New Jersey Maternal Mortality Dashboard website.

Figure 5:

Maternal mortality ratio (number of maternal deaths per 100,000 live births) for non-Hispanic White women (A) and percent difference in NTSV Cesarean rate from the New Jersey statewide 2020 goal of 24.7% (B) in 2017 in Cape May county, New Jersey.

Maps are screen shots from the New Jersey Maternal Mortality Dashboard website.

Discussion

Access to publicly available, accurate and detailed maternal mortality statistics in the United States is generally available but may not be easy to access. Moreover, interested parties may not be able to evaluate the range of variables that could be important to understanding maternal mortality. Further, multiple perspectives must be considered when interpreting data and seeking solutions (e.g., policy, economics, education, and integrated clinical support for physical and mental health). Meaningful maternal mortality data are not available for all states and are not standardized across states, leaving a gap in information [15]. Before implementation of the checkbox for pregnancy-related deaths and pregnancy-specific questions (e.g., timing of death in relation to pregnancy) on death certificates in 2003, maternal deaths in the NVSS were likely underreported [9, 10]. However, states implemented this information in a staggered fashion; therefore, evaluating trends in maternal mortality in the United States was incomplete between 2003 and 2017 [9, 10]. Use of the NJMMD can help fill existing gaps in information and our understanding of trends in maternal mortality for the state of New Jersey by providing detailed measures of maternal mortality, as well as factors that may be associated with the deaths. Further, as part of the ERASE MM program [20], the CDC offers tools, resources, and support to MMRCs to help improve review processes that inform recommendations for preventing future deaths.

To our knowledge, NJMMD is the first publicly available, interactive, state-focused maternal mortality tool that takes into account the intersection of social and demographic determinants of health (i.e., daily social, environmental, and economic conditions, such as where we live, work, and play), which are being increasingly recognized as playing important roles in health outcomes. Though some organizations and state governments have maternal morbidity and other maternal and child health statistics — sometimes in visual formats — on their websites [27], [28], [29], none has the interactive capability and scope of the NJMMD. We are aware of a similar interactive tool, the Maternal Mortality and Morbidity Interactive Dashboard (3MID), which was developed to evaluate the impact of social determinants of health on maternal mortality and serious illness (morbidity) [30]. However, this tool is not publicly available.

Considering intersectionality is important because it acknowledges that everyone has unique experiences of discrimination and oppression, and that anything and everything that can marginalize someone must be explored. For communities of color, social and demographic determinants (e.g., healthcare services and insurance, early access to care, housing and location [rural vs. urban], employment, income, and education) influence health as much as biologic determinants [31]. The NJMMD also takes into account the intersection of structural determinants of health (e.g., racial disparity, sexism, and classism), which drive and influence social determinants of health, and play a role in high maternal mortality rates in the United States [32, 33]. Although the precise reason(s) the maternal mortality rate is substantially higher in non-Hispanic Black and Native American women than non-Hispanic White or Hispanic women is unclear, a main reason for high maternal mortality rates in the United States compared with other developed nations is the disproportionately higher rates of pregnancy-related deaths in these populations [34]. Factors that contribute to negative outcomes in non-Hispanic Black pregnant women include structural racism, weathering, historical lack of trust in the health care system, racial bias among health care providers, victim blaming, limited diversity among providers, lack of access to or barriers to use of health care services, and higher rates of comorbid conditions such as diabetes, hypertension, and obesity [33], [34], [35], [36], [37]. Black women generally represent a marginalized population that is not sufficiently heard or trusted [34, 35, 38]. Moreover, when they speak up about maternal health issues, including during and/or after their pregnancy, their concerns are minimized or not taken seriously.

The NJMMD allows users to quickly see how demographic, social, structural, and public health factors could be correlated with maternal mortality, and how trends and patterns could help inform effective public health interventions. In addition, users can explore the code, data, and methodology behind the NJMMD and apply the same to the maternal mortality data for other states. Finally, the NJMMD could serve as an important research tool for students and scholars.

Although we provide a limited number of use cases as examples, the NJMMD can potentially be employed to study important questions related to women’s health outcomes in New Jersey, such as the impact of allocation or availability of certain resources (e.g., doulas, providers such as midwives and obstetrics/gynecologists, and social workers) on maternal mortality rates, or correlations between differences in maternal mortality rate among counties and contextual variables (e.g., education; employment; poverty; health insurance coverage; provider race; and presence of provider, particularly midwifery, deserts) available in NJMMD or other external sources.

Regardless of the complexity associated with understanding historical or new data on maternal mortality, the maternal mortality crisis must be a clear focus in the United States [13]. Recognition of the drivers, and identification and implementation of solutions, are urgently needed.

Of import is the understanding that some drivers and solutions for the maternal mortality crisis may vary from state to state, as is does from country to country, while other drivers and solutions may be common across geographies. For example, in a case-control study of women aged ≥35 years in the United Kingdom [39], five risk factors — inadequate use of antenatal care (adjusted odds ratio, 23.62 [95% CI, 8.79–63.45]), comorbidities (5.92 [3.56–9.86]), previous pregnancy complications (2.06 [1.23–3.45]), smoking (2.06 [1.13–3.75]), and age (1.12 [1.02–1.22]) — that were associated with maternal mortality are likely also applicable in the United States.

Finally, many of the social and structural determinants of health can be impacted by crisis situations, such as the COVID-19 pandemic [40]. Evidence indicates some of the drivers of maternal mortality were further exposed and exacerbated by the COVID-19 pandemic. This catastrophic event, which highlighted structural racism in the United States, led to the maternal mortality crisis taking on new urgency.

In addition to inherent strengths (e.g., simple, easy-to-use, and exportable, interactive presentation of existing, publicly available data that takes into account the intersection of social and demographic determinants of health; county-level data and analyses), the NJMMD also can be used to guide resource allocation and channel funding to areas of greatest need. However, although the NJMMD can help identify correlations and not causation, data could be used to inform evidence-based public health and policy interventions for improving maternal health in the state. Further, given relatively small sample sizes in some counties, trends and correlations for individual counties may not always be informative.

In conclusion, the United States is in the midst of a maternal mortality crisis, primarily because of high pregnancy-related deaths among non-Hispanic Black women. Some states, including New Jersey, have particularly high rates of maternal mortality. The reporting of data on maternal mortality and related lexicon are inconsistent across states and agencies. The NJMMD may fill a gap in public information on maternal mortality in New Jersey, as well as help identify how maternal mortality is related to social, social demographic, and structural determinants of health (e.g., age, race, access to health care, geography, and poverty).