Research and Professional Literature to Inform Practice, November/December 2025

Given significant problems facing maternity care providers and patients in the United States, this section provides an update on access,1 severe maternal morbidity,2, 3 and maternal mortality data.4, 5 Maternity care access is a problem attracting increasing attention, with significant variation in access across states.6 States that are more rural are experiencing greater losses in service coverage.7 Although rural areas have been the focus of much of the news on closing maternity units, these units are also more likely to close in areas serving Black and Brown women.8, 9 The current study charts the decline in maternity care access across the United States from 2010 to 2022 using the American Health Association Annual Survey data. The authors found that 7 states had a quarter or more of their hospitals lose maternity services by 2022. Five states had 25% or more of urban hospitals lose maternity services, and 12 states had the same loss among their rural hospitals. Unfortunately, Exhibit 1 in the article that presents these findings does not stand alone and requires the article text to be understandable. Although some states experienced no loss in maternity services (Delaware, Utah, and Vermont), other states had more than 25% of their hospitals close maternity services (Iowa, Oklahoma, Pennsylvania, Rhode Island, South Carolina, the District of Columbia, and West Virginia). As of 2022, 8 states offered maternity care in less than a third of their rural hospitals (Alabama, Florida, Illinois, Mississippi, Nevada, North Dakota, Oklahoma, and West Virginia). Exhibit 2 provides an excellent visual of the problem with a map of the United States in which the percentage of maternity care loss is color-coded and highly rural states are denoted with hatch marks. This allows readers to easily see the confluence of rurality and loss of maternity care services. The greatest losses are predominantly in the Midwest to South and in parts of Appalachia, although loss is evident across the country. Specifically, Florida, North Dakota, and West Virginia had the highest percentage of rural hospitals without maternity care, ranging from 87% to 70.4%. South Dakota and Hawaii had the greatest percentage of urban hospitals without maternity services (72.7% and 62.5%, respectively). Rural-urban differences in closures were most pronounced in Florida, New Hampshire, Pennsylvania, and South Carolina. In their discussion, the authors note that closures were an issue prior to 2010 but that although many groups are making efforts to address access, the problem continues to grow. They mention that differences in state policy environments likely shape access but do not specify particular policies. Given the diversity of states experiencing significant decreases in maternity care access, pinpointing policies contributing to the issue is unlikely to be straightforward. Although not mentioned in the article, consideration must be given to the economics of health care, which is related to but separate from policy. Finally, the authors note that these closures may concentrate care in larger hospitals better suited to high-acuity maternity care but do not mention associations with specific birth outcomes.8, 9 Continuing the examination of the relationship of access to birth outcomes, these authors looked at nontransfusion severe maternal morbidity (nt-SMM) in Louisiana from 2016 to 2021 and the associated sociodemographic characteristics.2 To conduct this birth cohort study, the authors used a statewide inpatient birth hospitalization discharge data set for the specified time period. The authors ran mixed-effects logistic regression models with the women nested in the ZIP codes where they lived. They defined nt-SMM according to the Centers for Disease Control and Prevention definition. Of the 326,597 births in the data set, 0.77% involved nt-SMM. After controlling for confounding factors, including medical comorbidity and sociodemographic characteristics, Black non-Hispanic and Hispanic women still had increased odds of nt-SMM compared with White non-Hispanic women (adjusted odds ratio [aOR] 1.36, 95% CI 1.23-1.51 and aOR 1.39, 95% CI 1.21-1.59, respectively). None of the associations between ZIP code poverty or maternity care desert status and nt-SMM were statistically significant, even while stratifying by race and ethnicity. In the discussion, the authors state that Black women and those living in the highest-poverty ZIP codes had the highest prevalence of nt-SMM. The greatest odds of nt-SMM, however, was associated with Hispanic ethnicity. They go on to claim that their results suggested that Black women in maternity care deserts experience greater odds of nt-SMM but then qualify the statement. The 95% CI includes the null, and therefore the findings should be viewed with caution. Indeed, technically, the finding should not be reported as suggestive at all. What the 95% CI that includes the null suggests is that the authors cannot be certain that their method would capture the true population interval 95% of the time. Although it seems likely that Black women living in maternity care deserts would experience worse care outcomes, it is critical that scientists are precise if the aim is for research to shape policy and practice. In this study,3 the authors examined the non-Hispanic Black-White gap in severe maternal morbidity (SMM) up to 42 days postpartum within Medicaid and privately insured births. To do this, they used hospital discharge abstracts from Georgia linked to birth and fetal death records from 2016 to 2020 and linked these with a number of other publicly available data sets (eg, the American Hospital Association Annual Survey). They identified 389,828 births covered by Medicaid or private insurance during the study time period. As with the above article, the authors used the Centers for Disease Control and Prevention's definition of SMM and excluded transfusion. In their models, they controlled for sociodemographic characteristics, medical and obstetric risk factors, access to care defined as third-trimester care, hospital characteristics, and residential characteristics. It appears that they did not use a validated algorithm to identify obstetric and medical risk but chose some of the most common risk factors (eg, diabetes, preeclampsia). They then used the Oaxaca-Blinder algebraic decomposition method to understand the difference in SMM between Black and White women. There were marked differences between the Black and White women in the study, from their sociodemographics to their health risks and where they gave birth. Black women, for instance, were more likely to give birth at hospitals with poorer nurse staffing, safety scores, and physician communication. The authors found that, for both insurance groups, hospital fixed-effects explained the greatest percentage of the discriminatory effect on the Black-White gap in SMM (47% within Medicaid and 14.1% within the privately insured). In the discussion, the authors interpret this finding as indicating that unobserved, within-hospital characteristics or processes in which Black women give birth may result in the SMM differences and that this is more marked in those hospitals where women insured by Medicaid give birth. In other words, it is the unmeasured processes and culture of care within hospitals, especially for individuals insured by Medicaid, that are significantly responsible for racial disparities in SMM. Sabr Y, Lisonkova S, Mayer C, Joseph KS. Temporal changes in the contribution of chronic disease to maternal mortality in the United States. Paediatr Perinat Epidemiol. 2025; 39(6):527-539. https://doi.org/10.1111/ppe.70034 World Health Organization. International Statistical Classification of Diseases and Related Health Problems. 10th Revision. 2nd ed. World Health Organization; 2004. https://apps.who.int/iris/bitstream/handle/10665/42980/9241546530_eng.pdf. The current study examines state-level maternal mortality trends from 2018 to 2023, including the contribution of COVID-19, using the National Vital Statistics System data. Maternal deaths were defined as those occurring within 42 days of a live birth or an abortion. The authors used hierarchical Bayesian models with spatiotemporal random effects to estimate smoothed MMRs (maternal deaths per 100,000 live births), as well as the percentage of maternal deaths involving COVID-19. The authors found that the MMR peaked across all states in 2021 and declined subsequently, and that these changes were almost entirely due to COVID-19. Put differently, if maternal deaths involving COVID-19 were excluded, the MMR would have been relatively stable throughout this period, both overall and across individual states. Although the trends were similar across states, some states had higher MMRs than others. For instance, California had the lowest MMR (9.4-12.6) and a smaller 2021 peak compared with other states (12.6), although it did have one of the highest percentages of maternal deaths involving COVID-19 in 2021 (58.1%). The authors do note, however, that California did not use the standard pregnancy checkbox until late 2021, so maternal deaths within 42 days could not be identified. Other states, meanwhile, had MMRs greater than 50 of 100,000 live births in 2021. If COVID-19–related deaths were excluded, the authors reported, then MMR rates were stable across most states. They did note, however, variation across states in both maternal mortality reporting and trends. This is an interesting article to review next to the preceding one, with one looking at overall and race- and ethnicity-specific MMR and the other examining state-level MMR. The 2 articles used different data sets and slightly different definitions and inclusion and exclusion criteria. The article by Rossen et al,5 for instance, included a second outcome and examined later maternal deaths, which were excluded from the article by Sabr et al.4 Maternal mortality calculation is based on definitions that are worth attention. The denominator is all live births in a period of time. The numerator, however, is defined as all deaths of women while pregnant or within 42 days of pregnancy termination1 (eg, live birth or abortion) from any cause related to or aggravated by the pregnancy or its management but not from accidental or incidental causes.2 In other words, although maternal deaths associated with an abortion are captured in the numerator, they are not captured in the denominator. Furthermore, later maternal deaths (deaths between 42 days and the end of the first year postpartum) are not included in this definition. Deaths with accidental injury, self-harm, or assault are also excluded. All of these caveats are problematic in one way or another, from missing maternal deaths that are related to or aggravated by pregnancy (eg, death due to intimate partner violence aggravated by pregnancy at any time, or death from multiorgan failure secondary to giving birth at 42 days postpartum) to counting deaths in the numerator that are not captured by the denominator (eg, when the end of pregnancy is an abortions or miscarriage). Maternal mortality calculations are estimates, and we should be attentive to who is included and excluded. In this section, an array of possible innovative solutions, spanning state1 and organizational policy,2 as well as a midwifery model of care,3 are reviewed. Connecticut implemented a deidentified notification policy for infants with prenatal substance exposure in 2019, becoming the first state to do so.1 Child Protective Service (CPS) notification, let alone maternal penalization for prenatal substance use, is a critical, and ethically fraught, aspect of providing care to women who use substances prenatally. The new policy does allow for traditional CPS notification with a safety screener that is integrated into the notification system and asks whether (1) the child has tested positive for substances due to maternal use; (2) there is a concern that maternal substance use will impact her functioning; and (3) there is a concern for abuse or neglect. The authors of this study focused on testing whether the deidentification policy resulted in an increase in CPS interactions. They used vital and child welfare records from March 2017 to July 2022 to examine the effect of the legislation. Specifically, they had data on screened-in CPS reports from the Department of Children and Families, as well as birth certificate records, and Connecticut Open Data. Bayesian interrupted time series models, which are used to measure policy effects, were used to estimate monthly reports to CPS and foster care placements. The average rate of prenatal substance use reports to CPS declined from 210 to 190 reports per 10,000 births after the policy change, with a concordant foster care placement due to substance use decrease from 54.5% to 43.1%. Over time, the policy was associated with larger and larger decrease in reports, from a 16% decrease at 12 months to a 51% decrease at 36 months after policy implementation. On a more granular level, the adjusted rate of reports per birth decreased by 7% per month after the policy was put in place, and foster care placements decreased by 4% per month. Although the authors’ findings suggest the legislation was associated with a decrease in CPS encounters, it is not clear whether this is an all-around benefit with improved maternal, infant, and family outcomes. For example, it would be helpful to see if the legislation was associated with increased care access for pregnant women with substance use disorder. The use of the screener, while allowing for traditional CPS notification if needed, also opens a door for bias in reporting, although this was not mentioned by the authors. Caseload midwifery, also called the continuity of care model, is the original midwifery model of one (or a few) midwives taking care of the same group of women, allowing deep relationships to build between midwife and patient. This is another article,2 in a long stream of articles, that has found caseload midwifery to be beneficial for patient, provider, and payor. In the current study, the authors conducted a descriptive qualitative study regarding the experiences of midwives providing a culturally tailored caseload model to First Nations women in Melbourne, Australia. The culturally tailored caseload model, Baggarrook Yurrongi (Women's Journey in Woiwurrung), involved cultural awareness training given to the midwives by the Aboriginal Community Health Organization or Aboriginal Health Unit staff, including information on the impact of colonization, trauma-informed practice, skills for implementing cultural safety, and respectful terminology, among other topics. This culturally tailored caseload model was provided to First Nations women and non-Indigenous women having First Nations newborns. All the midwives working in this model were invited to participate in this study that used semistructured interviews to explore their experiences. The model was implemented at 3 different sites, and midwives were invited to participate at several different time points, including 6 months after they started in their position, after 2 years in the job, when the project ended, or when they left their position. All the sites had preexisting caseload groups. Interviews were conducted in 2018 and 2021, and the authors used an inductive thematic approach for analysis. Twenty interviews were conducted. The primary theme that emerged from the work was that of “stronger connections for better outcomes,” which had 4 subthemes: (1) strengthened connections between the midwife and woman; (2) strengthened connections to navigate services and systems; (3) strong connections among caseload midwives; and (4) responsive systems and management required for strong connections and sustainability. For the first subtheme, midwives noted that, because of the relationship nurtured in the model, women did not need to keep repeating their stories, and midwives were able to tailor care for the individual woman. Women at any level of risk were able to participate in this model, and midwives found that they were better able to support and advocate for women, from texting between appointments, to coordinating care, and being flexible with appointment times, which led to improved access and satisfaction with care. Bridging the first 2 subthemes, midwives reported a better understanding of the mistrust in the healthcare system experienced by First Nations people, reinforced by more than half of the midwives seeing or experiencing racism toward this group from other hospital staff. The midwives, who all had experience in the caseload model, noted how this version was different, with more administrative tasks, coordination, and follow up needed to ensure continued care. One site offered “reflective supervision” with a hospital social worker that the midwives noted to be extremely valuable in supporting them in caring for women in the caseload model. In the final subtheme, midwives noted persistent organizational barriers for the sustainability of the model. These included a need for organizational support and resources (eg, increased staffing, adequate clinical space, dedicated obstetric and social work staff), as well as a system focus on building the First Nations midwifery workforce. Overall, this work is suggestive of both the transformative power and significant resources needed for this excellent model of midwifery care. In this final section, a critical history lesson with multiple problems and the solutions that were proposed is offered, for better or worse.1 It provides an important reminder of the double-edged nature of legislation and policy. This article1 provides a nuanced view of public health initiatives in West Virginia in the late 19th and early 20th centuries. These were meant, ostensibly, to bring resources and attention to maternal-infant health but led to the decline of traditional midwifery through regulatory processes and birth medicalization. The author provides a fascinating account of the West Virginia context. While West Virginia coal helped drive industrial development in the United States, coal companies would fund basic medical care from miners’ pay—and this care explicitly excluded pregnancy care. As a result, pregnancy and birth care were provided by local midwives and family members. The first Board of Health was established in West Virginia in 1881 and was focused on unlicensed doctors and folk healers, not midwives. The focus on midwives came almost 3 decades later, with the transition to a state-run Department of Health. At this time (and coinciding with a broader movement against midwives), the Board to against midwives who were not their (eg, without or The author that medical to midwifery as a problem to be to birth outcomes, them for outcomes, and that this cause was also by White women. a different public health noted a of maternity care in the state and for the state to services. It is interesting to note that the which provided to states to health services and had severe for traditional midwives, including midwives in the was by the American Association as The author that this was a result of a power between and as the state health were by public health and and it was women who to the the years after the was the to the to this, however, West Virginia the and the and used it to of traditional midwives in maternity care. Although were not it was that maternal and deaths were due to midwifery care, although midwives of the births in the state at that time. from to examine midwives to them These did not go as the public their midwives to and even the Health Department that midwives provided critical care, with no other in the In West Virginia a and midwives for the first time, which also required physician physician midwifery included and The result of the was the decline of traditional midwives in the to to midwives, which was included in the were found because of the of the The came to an end in coinciding with the the end of the public health from to and for traditional midwives The view that birth was the of the medical and was It was not until that a was allowing alone to in West to allow midwives to in West Virginia was proposed in and did not