Introduction

Deliveries involving severe maternal morbidity (SMM), which is defined as unexpected outcomes of labor and delivery that result in significant short- or long-term consequences to a person’s health, continue to rise in the United States (US) [1]. It is estimated that 50,000–60,000 pregnant people in the US experience SMM [1,2]. SMM includes life-threatening conditions, such as acute myocardial infarction, acute renal failure, amniotic fluid embolism, disseminated intravascular coagulation, or sepsis. Prior research indicated a 75% increase in severe complications during delivery hospitalizations and a 114% increase in severe morbidity during postpartum hospitalizations between the 1998–1999 and 2008–2009 periods [3]. The Centers for Disease Control and Prevention (CDC) reported close to a 200% increase from 49.5 in 1993 to 144.0 per 10,000 deliveries in 2014; this increase was mostly attributed to blood transfusions, which increased from 24.5 in 1993 to 122.3 in 2014 [1].

We described a similar increase in SMM rates at the time of delivery in the Commonwealth of Massachusetts from 129.4 in 2009 to 214.3 per 10,000 in 2018 [4]. One of the Massachusetts Title V program priorities is to eliminate inequities in SMM. An Act to Reduce Racial Inequities in Maternal Health [5] was signed in 2021 in Massachusetts and established a commission to investigate and study ways to reduce or eliminate racial inequities in maternal mortality and SMM in the Commonwealth using evidence-based, best or promising practices. A large body of literature has documented racial/ethnic inequities in reproductive health [6–10] and racism as a fundamental cause of such inequities [11]. A systematic review of 37 studies, mostly conducted in the US and with physicians, found statistically significant evidence of racist beliefs, emotions, or practices among healthcare providers in relation to minority groups in 26 of these studies [12].

A population-based California study found Black non-Hispanics having 44% higher rates of SMM than non-Hispanic Whites, even after adjustment for demographic and medical risk factors [11]. A retrospective study in a single hospital in Illinois that reviewed selected SMM cases indicated that compared to non-Hispanic White, non-Hispanic Black birthing people were more likely to experience a disproportionately high burden of SMM due to preeclampsia and eclampsia (31% vs 18.1%) and were more likely to need improvement in care compared with non-Hispanic White birthing people (53% vs 39.0%) [13]. Black birthing people have not historically received optimal care and racism as a contributor to racial inequities in healthcare is now well recognized [9,14–17].

Globally, where SMM reviews were conducted, the greatest preventable factors identified were provider-related, specifically the lack of identification of “high risk” status, and delays in both diagnosis and treatment [18,19]. In a study that examined preventability of maternal death, near miss, and severe morbidity, provider-related factors were mentioned for approximately 90% of the preventability in all three groups, while system factors were cited in 33% to 47% of preventable events, and patient factors were only reported in 13% to 20% of preventable cases. The major preventability factors at any point in time in the morbidity and mortality continuum were inappropriate and incomplete management, including failure to diagnose, delays in diagnosis, poor communication, and inappropriate referrals [19].

While reducing maternal death and SMM was a Healthy People (HP) 2010 and 2020 goal and remains one of the goals for HP 2030, limited progress has been made in achieving these goals [20], with the release of the 2020 maternal mortality rates reporting the highest US rate since 1968 [21]. To identify factors contributing to the persistent and rising rates of SMM and racial/ethnic inequities, we used a longitudinally linked population-based data system in Massachusetts to examine SMM trends over 20 years, the Black non-Hispanic/White non-Hispanic rate ratio for each SMM condition, and factors associated with SMM by race/ethnicity.

Materials and methods

Study sample

This analysis included in-state deliveries from January 1, 1998 to December 31, 2018 (N = 1,539,137) to 948,643 Massachusetts residents resulting in 1,573,812 infants. After we excluded delivery records that did not link to a hospital record, early losses, and ectopic and molar pregnancies (n = 17,148), there were 1,521,989 deliveries in our study sample, representing 1,010,757 deliveries to White non-Hispanic, 228,812 to Hispanic, 133,224 to Black non-Hispanic, and 149,196 to all other groups. When examining the factors associated with SMM, we focused our analyses on White non-Hispanic (White) and Black non-Hispanic (Black) birthing people. Deliveries with missing data on maternal age, education, insurance, or parity were excluded. To assess the effects of body mass index (BMI) on SMM, we further restricted our study sample to deliveries in 2011–2018 when the BMI data were available (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flow diagram of study sample with inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0279161.g001

Results

This study included 1,521,989 deliveries to Massachusetts residents during 1998–2018. A substantial number of these deliveries (66.4%) were to White birthing individuals, followed by Hispanic (15.0%), Black (8.8%), and Other non-Hispanic (9.8%). Between the 1998–2000 and 2016–2018 periods, overall SMM20 rates increased significantly from 36.9 to 92.3 per 10,000 deliveries. While the rates of SMM20 during the same periods steadily increased for all racial/ethnic groups, Black birthing people experienced the greatest absolute increase going from 69.4 in 1998–2000 to 173.7 per 10,000 deliveries in 2016–2018. The rates of SMM20 per 10,000 deliveries increased from 33.4 to 76.1 for White, 39.2 to 92.8 for Hispanic, and 37.9 to 99.6 for Other non-Hispanic birthing people between 1998–2000 and 2016–2018 (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Severe maternal morbidity^a by race/ethnicity using three-year rolling average: Massachusetts, 1998–2018.

^aSevere maternal mortality without blood transfusion (SMM20).

https://doi.org/10.1371/journal.pone.0279161.g002

Most notably, the rate for Black birthing people in 2016–2018 (173.7 for every 10,000 deliveries) was 128.3% higher than the rate for White, 87.2% and 74.4% higher than the rates for Hispanic and Other non-Hispanic, respectively. The gap between Black and White birthing people in the earlier period (1998–2000) was two-fold. By 2016–2018, the SMM rates for Black birthing people were 2.3 times higher than that of White. Hence, we chose to focus our analysis on better understanding the factors that contribute to SMM in the two racial/ethnic groups with the highest and lowest SMM20 rates. Fig 3 presents the Joinpoint analysis illustrating that SMM20 rates for Black birthing people increased significantly between 1998–2000 and 2002–2004 with an annual percentage change (APC) of 10.6%, declined between 2003–2005 and 2007–2009 with an APC of 5.2%, then steadily increased with an APC of 8.1% and 12.3% between 2008–2010 and 2016–2018. For White birthing people, the rates increased steadily from a lower base from 1998–2000 to 2011–2013 with an APC of 3.1%, and a more pronounced increase between 2012–2014 and 2016–2018 with an APC of 8.6%.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Severe maternal morbidity^a for Black and White birthing people: Massachusetts, 1998–2018.

^aSevere maternal mortality without blood transfusion (SMM20).

https://doi.org/10.1371/journal.pone.0279161.g003

In examining SMM rates for each condition by race/ethnicity, we found that Black birthing people had higher rates for every individual condition (including transfusion in this comparison) compared to White birthing people. For temporary tracheostomy, sepsis, acute renal failure, and pulmonary edema, and sickle cell anemia, the Black/White rate ratios ranged from 3.30 to 102.4. For transfusion, eclampsia, adult respiratory distress syndrome, ventilation, shock, conversion of cardiac rhythm, air and thrombotic embolism, cardiac arrest, and amniotic fluid embolism, the Black/White rate ratios ranged from 2.00 to 2.97 (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Severe maternal morbidity^a indicators for White non-Hispanic and Black non-Hispanic birthing people: Massachusetts January 1, 1998-December 31, 2018.

https://doi.org/10.1371/journal.pone.0279161.t001

Our stratified analyses using GEE for Black and White are shown in Table 2. Compared to birthing people aged 25–29 (reference group), there are no significant differences in SMM rates among birthing people aged less than 18 years among Whites [aRR 1.01 (95% CI: 0.75, 1.37)] and Blacks [aRR 0.91 (95% CI: 0.62, 1.32)]. However, among birthing people ages 18–24, SMM rates were significantly lower compared the reference group among both Whites [aRR 0.75 (95% CI: 0.67, 0.85)] and Blacks [aRR 0.77 (95% CI: 0.65, 0.92)]. Among White birthing people aged 30 years and older, SMM rates significantly increased with increasing age across all age categories. While among Black birthing people SMM rates show a similar dose response with age, the rates were not significant for the 30–34 age group [aRR 1.10 (95% CI: 0.95–1.28)].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics associated with severe maternal morbidity^a for White non-Hispanic and Black non-Hispanic birthing people: Massachusetts, January 1, 1998-December 31, 2018.

https://doi.org/10.1371/journal.pone.0279161.t002

While education level was not significantly related to SMM20 for Blacks, high school or less was significantly associated with higher SMM rates among Whites [aRR 1.23 (95% CI: 1.13, 1.34)]. SMM rates for those on public insurance were significantly higher with comparable risks for White [aRR 1.23 (95% CI: 1.13, 1.33)] and Black [aRR 1.25 (95% CI: 1.09, 1.44)] birthing people. Except for 2001–2003 and 2004–2006 periods, SMM rates increased significantly overtime among Whites, while for Blacks, SMM rates increased significantly only between 2010 and 2018. While having multiple births was significantly associated with the risk of SMM for both groups, the risk was more pronounced among Whites [aRR 3.01 (95% CI: 2.71, 3.34)] than Black birthing individuals [aRR 2.25 (95% CI: 1.78, 2.85)]. History of hypertension was significantly associated with higher risk of SMM among Blacks [aRR 2.16 (95% CI: 1.88, 2.49)] and Whites [aRR 1.81 (95% CI: 1.65, 2.00)]. History of diabetes was also associated with higher risk of SMM among Black and White individuals but is more pronounced in Blacks [aRR 1.45 (95% CI: 1.17, 1.81)] compared to Whites [aRR 1.23 (95% CI: 1.05, 1.44)].

Adequacy of prenatal care also showed similar patterns for Blacks and Whites, where inadequate and intermediate inadequate care were significantly associated with higher rates of SMM. The rates of SMM decreased with increasing parity and decreasing facility number of annual deliveries. When we replicated our stratified analyses for White and Black to include BMI in the model for the years for which it was available, we found that BMI was not significantly associated with SMM among Black birthing people, but BMI ≥30 (obese) was significantly associated with SMM [aRR 1.18 (95% CI: 1.02, 1.36)] among White birthing people (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics associated with severe maternal morbidity^a including body mass index for White non-Hispanic and Black non-Hispanic birthing people: Massachusetts February 1, 2011-December 31, 2018.

https://doi.org/10.1371/journal.pone.0279161.t003

Discussion

Our study shows that while the rates of SMM for all race/ethnic groups have continued to rise over the last two decades, Black birthing people have persistently experienced the highest rates of any race/ethnic group, and those rates increased by 150% between the 1998–2000 and 2016–2018 periods, widening an already large Black/White difference. This finding is consistent with other studies. [26,27] Our study also found that for every SMM condition, Black birthing people had much higher rates of SMM, with gaps between Black and White, ranging from a low of 1.11 for heart failure during surgery to a high of 102.4 for sickle cell anemia. Prior research suggests that while Black birthing people are more likely to develop these conditions, they are less likely to have their conditions adequately managed, and more likely to have complications and mortality from these conditions [28]. In our adjusted model, in additional to being Black, factors significantly associated with SMM included multiple birth, history of hypertension, history of diabetes, being on public insurance, having a high school or less than high school education, and inadequate prenatal care. Facility number of annual deliveries, age, parity, and year of delivery were also associated with increased risk of SMM for Black birthing people, suggesting a dose response.

Our stratified analyses sequentially adjusting for covariates showed that education (high school or less) was significantly associated with SMM among Whites, while the risk of SMM among those who are on public insurance was comparable among White and Black birthing people. The aRR for multiple births was higher among White people, while the aRRs for history of hypertension and diabetes were higher among Black birthing people. When we include BMI in the model, we found that BMI was not significantly associated with SMM among Black birthing people, but BMI ≥30 (obese) was significantly associated with SMM among White birthing people.

Most SMM conditions are interrelated, preventable, and may be related to preexisting conditions or develop secondary to diseases of pregnancy [29]. For example, since eclampsia is complication of severe preeclampsia, understanding and treating preeclampsia can prevent SMM due to eclampsia and mortality [30,31].

Our study showed that the Black/White rate ratio for eclampsia was 2.13, which is consistent with findings from the CDC that reported a 3.1 times increased incidence of preeclampsia/eclampsia in Black patients compared to White patients [28]. These findings are disturbing since treatment for preeclampsia and eclampsia is available even in developing countries. Magnesium sulfate (MgSO4), the drug of choice for treating and preventing eclampsia [32–34], has been recognized by the World Health Organization (WHO) as the most efficient and safest medication for the treatment of preeclampsia and eclampsia; WHO has made it available on the Essential Medicines List for this explicit use. Findings from clinical trials conducted between 1995 and 2002, indicated that MgSO4 was the most effective when compared with other treatments, including diazepam and phenytoin. Patients treated with MgSO4 had a 52% lower recurrence of eclampsia than those treated with diazepam and 67% lower recurrence of convulsion than those treated with phenytoin [32]. MgSO4 was also found to decrease the occurrence of eclampsia by more than 50% and maternal deaths by 46% [35]. In addition, MgSO4 is cost-effective and approximately costs $0.10/ml [33].

Our study also found that Black birthing individuals experienced a higher prevalence of acute renal failure (20.8/10,000 deliveries) compared to Whites (5.9/10,000 deliveries). While acute renal failure is a rare event, it can be associated with substantial morbidity and mortality. The etiology of renal failure is complex and could include prerenal factors, which lead to decreased renal perfusion, intrarenal factors which affect the renal parenchyma, or postrenal usually due to obstructions. Prerenal causes of acute renal failure could result from hypovolemia due to obstetric hemorrhage (abortion, placenta previa, placental abruption, uterine rupture or postpartum hemorrhage), but may also arise from sepsis and severe cases of hyperemesis gravidarum, leading to ischemia from decreased renal perfusion and hypotension, and from amniotic fluid embolism leading to disseminated intravascular coagulation (DIC), cardiac dysfunction, and hemorrhage causing intravascular volume depletion and reduced renal perfusion. About 40% of acute renal failure in pregnancy is caused by severe preeclampsia and hemolysis, elevated liver enzymes, and low platelet count (HELLP) syndrome, which can be prevented with adequate and timely management of the pregnant patient [36].

Our study shows that compared to Whites, for every 10,000 deliveries, Black birthing people were more likely to have a diagnosis of DIC (28.2 vs. 16.8), pulmonary edema (11.9 vs. 3.6), and sepsis (9.8 vs. 2.6). While sepsis is a common cause of DIC, other leading causes of DIC include preeclampsia, HELLP syndrome, placental abruption and postpartum hemorrhage [37]. An observational survey conducted in Japan showed that among 1,895 patients with sepsis who were treated in intensive care units, 29% were diagnosed with sepsis-induced DIC [37]. Hence, adequate and timely management of conditions such as sepsis, preeclampsia, HELLP syndrome, and postpartum hemorrhage is key to preventing DIC. Other prevention efforts for SMM should also be placed in the context of public health interventions.

A study indicated that late-season influenza infection between April and June, is associated with a higher risk of SMM and sepsis in pregnant patients and recommends providers remain vigilant as vaccination, early identification and treatment of influenza is associated with improved outcomes among pregnant people [38]. Other studies of pregnant patients with influenza documented increased risk of SMM, increased incidence of hospitalization [39], ICU admission, pneumonia, ventilator support, and maternal or fetal death [40,41], which could all be prevented via screenings and vaccination. A Massachusetts study indicated vaccination rates were significantly higher among pregnant patients whose provider offered or recommended the seasonal (75.8%) and pH1N1 (68.1%) vaccines compared with those who did not receive a recommendation (32.4% and 8.6%, respectively) [41].

Prior research has documented concerning patterns of delay in recognition of hemorrhage, hypertensive crisis, sepsis, venous thromboembolism, and heart failure [42]. Utilization of evidence-based protocols, triggers, bundles, and checklists can improve timely diagnosis and treatment and prevent or minimize the severity of illness, but also enable interdisciplinary, patient-centered care [43]. To facilitate timely recognition, diagnosis and treatment for patients who advance to critical illness, early-warning systems have been proposed by a multidisciplinary working group convened by the National Partnership for Maternal Safety. The Maternal Early Warning Criteria includes a list of abnormal parameters that indicate the need for urgent evaluation by a clinician with the ability to escalate care as needed [42].

It is also documented that even at comparable levels of access to care, people of color experienced a lower quality of health services and are less likely to receive even routine medical procedures than White Americans [44]. This is reflected in our findings showing that Blacks had higher rates for every SMM condition, which suggests a lack of timely diagnosis, failure to treat, poorer quality medical care due to racism or a combination of these. Black people have been found to be substantially more vulnerable to receiving poor quality of care [44]. Provider held unconscious bias was identified by the National Academy of Medicine as one of the three root causes of racial ethnic inequities; the other two being limited access to care and lack of trust in the healthcare system [9]. Providers’ bias in clinical encounters, has led to people of color receiving inferior medical care and fewer procedures compared to Whites across nearly every type of diagnostic and treatment interventions [31]. Black, Hispanic, and Native birthing people also report higher rates of mistreatment in perinatal care settings [45–47].

Our study has several limitations that should be noted. We relied on linked vital statistics and administrative data and these may not always capture the nuances of severe morbidities, though the algorithm used for measuring SMM has been subject to considerable analysis and testing [3]. Also, our study may not be generalizable to other states or the US because the Massachusetts birthing population is not representative of the nation, with a smaller proportion of births to Blacks (10.1% vs 15.0% in U.S.) than the national average. Our sample also has a larger proportion of births to older birthing individuals (57% age 30+) than the U.S. (42% 30+) [4]. Our study also may not be comparable to other SMM analyses since it does not include transfusion, though we think the greater reliability of measurement associated with the SMM20 justifies that limitation [22]. Finally, we included only delivery hospitalizations and did not capture birthing people hospitalized prenatally or postpartum to assess for SMM conditions that were not also present at delivery. We also included limited facility characteristics in this analysis although key institutional differences in where Black and White birthing people deliver have been previously reported [48]. A recent published study found that by using a population-based, longitudinally linked dataset to examine prenatal and postpartum hospitalizations, an additional 22% of cases of SMM between 2009 and 2018 were identified [4].

Our study has several strengths. First, it relies on a large population-based dataset that longitudinally linked birth and fetal death certificates with hospital discharge data, which allows for better identification of deliveries, since hospital discharge data do not have date of delivery and rely on ICD and procedure codes to identify deliveries. Our study relied on a robust linkage; data have been linked for 98% of live births and fetal deaths in Massachusetts for birthing individuals and their children since 1998. Our study was also based on two decades worth of data and allowed us to identify persistent trends. We conducted both adjusted and stratified analyses.

The fact that the rates for White non-Hispanic and all other race/ethnic groups were largely similar during our study period and continued to show similar patterns over time, while the rates and the gaps have continued to worsen for Black birthing people, indicates that Black birthing people have apparently not benefited from improved medical knowledge and care. It has been long recognized that race itself is not a risk factor; exposure to racism is the risk factor as it leads to discriminatory beliefs and behaviors toward Black birthing people [49]. In their review, Diop et al demonstrated the importance of addressing experiences of bias and racism directly with patients and the unique role healthcare providers can play. They proposed a framework of trauma-informed care, structural competency, provider bias and intersectionality when discussing patients’ experiences of racism [50]. Hardeman et al, have proposed to couple critical race theory with the previously described relationship centered care to improve the clinical experiences of Black birthing people [51,52].

An unbiased understanding of how SMM has affected different race ethnicity groups is key to improving maternal health and preventing SMM and mortality among Black pregnant people. SMM needs to be addressed as both a medical and public health challenge. This comprises identification of not only clinical but also social drivers of maternal morbidity, including racism, and other modifiable factors [31]. It is now recognized that an inpatient- or delivery- focused response alone is insufficient to adequately recognize, prevent or respond to significant morbidity during the antepartum period. To eliminate inequities, Carmichael et al. conceptualized SMM within both multidimensional structural and societal factors and the described pathways from specific clinical precursors to specific SMM indicators that can potentially prevent acute progression of disease to life-threatening conditions [31].

In a prior study, we suggested a need to ensure outpatient implementation and surveillance of care quality activities to identify and prevent morbidity from severe hypertension, sepsis and preeclampsia [4]. It is time for healthcare providers at all levels to acknowledge the historical legacy of racism and build a culture that centers on equity. State perinatal quality collaboratives should work with state health departments, Medicaid agencies, and community members with lived experience to better understand the degree to which implicit bias, racism, and discrimination affect Black birthing people at the community and institutional level and identify potential solutions.

Supporting information

Acknowledgments

We thank Mr. Qi Yu, MPH, PELL Programmer at the Massachusetts Department of Public Health for his contribution in the development of the PELL data system. We also thank Ms. Judith Jeanty, Project Manager, and Ms. Kayla J Kinsler, Research Assistant, Department of Public Health and Community Medicine, Tufts University School of Medicine for their contributions.