Stay Informed

Addressing The Maternal Health Crisis Through Improved Data Infrastructure: Guiding Principles For Progress

August 11, 2021
Health Affairs Blog
By Andre Chappel, Nancy DeLew, Violanda Grigorescu, Scott R. Smith

The United States has the highest maternal mortality rate among developed countries and experiences substantial disparities in maternal health outcomes, particularly by race and ethnicity. The most recent national report on maternal mortality in the US reveals a significant increase in the national maternal mortality rate in 2019 compared to 2018 (20.1 and 17.4 deaths per 100,000 live births respectively), with the rate among non-Hispanic Black women 2.5 times the rate among non-Hispanic White women and 3.5 times the rate among Hispanic women. Differences by age groups were also noted (12.6 deaths per 100,000 live births among women younger than age 25, 19.9 among those ages 25–39, and 75.5 among those ages 40 and older). A prior report using older and 10-year (2007–2016) aggregated data on pregnancy-related deaths (measured differently than maternal mortality) from the Centers for Disease Control and Prevention’s (CDC) Pregnancy Mortality Surveillance System indicate that American Indian and Alaska Native women also experience high mortality (2.3 times higher pregnancy-related mortality ratio compared to their non-Hispanic White, Hispanic, and Asian/Pacific Islander counterparts, and the gap widened among older age groups).

As a nation, we can and should provide better, more patient-centered care for expecting and new mothers that is responsive to their expressed needs—both to improve health outcomes for all mothers and their children and to eliminate current disparities.

Achieving Optimal Maternal Health For All In The US

Improving maternal health outcomes is an ongoing priority for the Department of Health and Human Services (HHS). HHS released a maternal health action plan in December 2020 with the vision of making the US one of the safest countries in the world for women to give birth. More recently, HHS marked Black Maternal Health Week by announcing actions to expand access to continuous health care coverage. The Centers for Medicare and Medicaid Services (CMS) approved section 1115(a)(2) Medicaid demonstration projects related to postpartum coverage for states such as Illinois, which became the first state to expand postpartum coverage for up to 12 months to Medicaid-eligible women; Georgia, which expanded postpartum coverage up to six months; and Missouri, which provides limited benefits up to 12 months postpartum for substance use disorder and mental health services. CMS will continue working with states that express interest in pursuing this option.

Furthermore, section 9812 of the American Rescue Plan Act authorizes all states to extend Medicaid eligibility through 12 full months postpartum, for a five-year period beginning April 1, 2022. HHS also announced a Notice of Funding Opportunity that will make $12 million available over four years for the Rural Maternity and Obstetrics Management Strategies program allowing awardees to test models that address unmet needs of populations that have historically suffered from poorer health outcomes, health disparities, and other inequities.

Progress in improving maternal health and decreasing rates of maternal mortality require advancements in prevention, treatment, and addressing social determinants of health (SDOH). A recent letter from CMS to state health officials described opportunities to better address SDOH under Medicaid and the Children’s Health Insurance Program, and to support states with designing programs, benefits, and services that can more effectively improve population health, reduce disability, and lower overall health care costs.

Research efforts are essential to identify the structural inequities and root causes of poor outcomes and to develop evidence-based interventions to address these causes. Having access to timely and high-quality data is a key component for advancing such research and monitoring progress. Specifically, making improvements to the nation’s maternal health data infrastructure—how data are collected, managed, and made accessible to researchers, as well as the capacity of researchers to analyze them—can inform future efforts and assessments of whether progress is being achieved. This is why an important pillar of HHS’ action plan is bolstering maternal health data and research.

portfolio of intradepartmental projects coordinated by HHS’ Assistant Secretary for Planning and Evaluation specializes in building this improved data foundation. Designed to build data capacity for conducting patient-centered outcomes research (PCOR), the Office of the Secretary Patient-Centered Outcomes Research Trust Fund (OS-PCORTF) was created by the Affordable Care Act in 2010 and was reauthorized for another decade in 2019 with maternal health research as one of its priorities. PCOR holds potential to reduce disparities in maternal health outcomes, as it generates evidence on the effectiveness of maternal health interventions and treatments and takes into consideration all relevant factors across the life course. Work from the OS-PCORTF reveals challenges and opportunities for advancing data infrastructure to improve maternal health, at the state and national levels.

Review Of State Initiatives Reveals Challenges In Data Collection And Data Linking

For several decades, states have been at the frontline in efforts to address the root causes of maternal deaths and implement strategies to prevent poor maternal outcomes. These initiatives create new collaborations among state professionals and organizations including medical providers, hospital administrators, public health officials, Medicaid officials, data specialists, researchers, and patient and community representatives. Critical to these efforts is the need for timely, high-quality data on the circumstances surrounding poor outcomes and to assess the effectiveness of interventions designed to improve care delivery. OS-PCORTF recently conducted semi-structured discussions with Medicaid medical directors, maternal mortality review committees, perinatal quality collaboratives, researchers, and data stewards across 21 states to better understand challenges in using maternal health data for research and quality improvement and to gather multistakeholder perspectives on state data capacity, timeliness, and quality.

These discussions surfaced three key challenges states are facing in leveraging data to improve maternal health outcomes. The first is incomplete and inconsistent collection of race, ethnicity, and other self-reported demographic data. All states described better collection of race and ethnicity data as a critical need to reduce health disparities, and most states acknowledged vital records as the most reliable source. Several states noted efforts to improve demographic data. For example, Massachusetts’ Medicaid program accesses hospital discharge data to extract race and ethnicity information as a reference when calculating rates of severe maternal morbidity.

Secondly, stakeholders noted the complexity of obtaining and linking data, citing legal and ethical considerations, limited resources to navigate data sharing agreements, and disparate timeliness of data sets. As an example, a discussant explained that epidemiologists at their state’s Department of Public Health (DPH) were required to complete separate data use agreements (DUAs) to review linked birth certificates with Medicaid claims, despite the fact that the state’s Medicaid and DPH agencies have a formal interagency DUA. This results in an onerous process for both the DPH data requestor and Medicaid as the claims data owner. Streamlining processes for sharing maternal health data across agencies is an important opportunity to improve maternal health data infrastructure that is within control of state agencies.

Finally, discussants noted that maternal mortality analyses have traditionally focused on obstetric complication measures, rather than other key drivers of poor maternal health outcomes including SDOH, behavioral health and substance use disorder, and chronic health conditions such as hypertension. However, chronic diseases and social risks are not always captured in clinical records proximate to the death. Most discussants mentioned that being able to acquire such data would help in identifying other medical and social risk factors that could be addressed through prevention including better linkages with social and community-based services.

Federal Support For States Linking Clinical And Other Data For Research

While much of the data needed to improve maternal health is collected at the state and local level, federal support is critical in helping to build the data infrastructure for research. Several OS-PCORTF projects are aimed at improving maternal health, but four in particular address some of the main challenges identified in our state stakeholder interviews.

The first, a project led by the CDC and conducted in partnership with the Substance Abuse and Mental Health Services Administration and CMS, establishes standardized data extraction from electronic medical records for analysis to inform person-centered care for pregnant women with opioid use disorder and their offspring. Through a network surveillance system called the Maternal and Infant Network, the project will monitor more than 2,000 mothers and their children through six years of age, using data collected from geographically diverse clinical sites.

Secondly, another CDC project is addressing data linkage challenges highlighted by state stakeholders. The project will link the Pregnancy Risk Assessment Monitoring System (PRAMS) with birth certificates and Medicaid claims and other clinical outcomes data. PRAMS is the only national surveillance system that provides data about mothers before and during pregnancy and the first few months following births, including information on their infants. PRAMS also captures self-reported information on behavioral health, social context (for example, intimate partner violence, housing insecurity, and incarceration), and other SDOH. This project will create a publicly available protocol outlining the data linkage process and how linked data can be accessed. In addition, a learning community will be established to provide technical support to researchers using the project’s data and its standardized methodology for linking data sets.

In a third project, teams at the CDC and the Health Resources and Services Administration are collecting electronic health record data from federally qualified health centers (FQHCs) to potentially link to outside data sources such as the National Death Index and Department of Housing and Urban Development administrative data. This data linkage would provide critical information on SDOH and mortality post-FQHC visits.

Finally, the National Institutes of Health and the Office of the National Coordinator, in partnership with the CDC, are working to harmonize both maternal and infant health data on medical conditions and interventions so that the data can be accessed through a FHIR® (Fast Healthcare Interoperability Resources) application programming interface. This will enable researchers to identify the full range of risk factors, including longitudinal medical history and basic socioeconomic and demographic characteristics that affect pregnancy outcomes for both the mother and the infant.

Principles For Moving Forward

Combining the key findings from our state-level maternal health interviews, the priority actions around data quality and timeliness in the HHS action plan, and the current efforts of the OS-PCORTF projects, we suggest the following principles to advance PCOR data infrastructure for maternal health efforts:

Recognize The Value Of Data Infrastructure For PCOR That Informs Decisions To Improve Maternal Health

Fostering a holistic person-centered approach must move beyond focusing on immediate symptoms to address underlying drivers of health. This approach requires supporting partnerships among individuals seeking care and all of those involved in providing services and addressing their needs. To understand how to effectuate such change, PCOR data should ideally capture a robust picture of not only personal characteristics of individuals and their clinical conditions and outcomes, but also underlying SDOH contributing to such outcomes. Information on clinical conditions and outcomes informs where clinical quality improvements can be made. Information on personal characteristics and SDOH sheds light on disparities and non-clinical interventions. For example, supporting efforts to coordinate referrals to social service providers and community-based organizations can be helpful in overcoming structural, including socioeconomic, barriers to good health outcomes.

Prioritize Efforts To Bolster Data Infrastructure Including Data Collection, Linkage, And Data Sharing To Support PCOR

The OS-PCORTF-funded projects highlighted in this blog post are designed to advance the linkage of clinical data with other data sources for maternal health PCOR; however, more work must be done to improve processes for data collection, linkage, and sharing. The ability to improve maternal health outcomes is highly dependent on having access to timely, high-quality data from a variety of sources. As state stakeholders also noted in the interviews, navigation of cross-sector data sharing can be complex and resource-intensive. Identifying sustainable strategies to consolidate data sharing and linkages, while maintaining rigorous data privacy and security standards, is an essential next step.

To that end, there are opportunities to foster cross-sector data sharing through community-level initiatives. OS-PCORTF recently examined community-level efforts to address SDOH through interviews with participants in three community-level initiatives. Their experience underscored the importance of cross-sector collaboration between state and federal governments including to develop data infrastructure.

In parallel, efforts are underway to expand access to and use of de-identified data sets for PCOR, priorities noted by a range of PCOR stakeholders who were asked to identify challenges and improvements for PCOR data infrastructure. Improved access to de-identified data resources—especially linked data resources that can inform person-centered care—will provide researchers critical information to facilitate clinical and policy analyses to address the maternal health crisis, while maintaining patient data privacy and security.

Minimize Data And Research Silos And Expand Collaboration Across Stakeholders

Making meaningful progress to combat the maternal mortality and morbidity crisis will require vision, investments, and commitment from providers, payers, patients, and other important stakeholders to advance PCOR data infrastructure. HHS will continue to foster partnerships between state and local governments and the many stakeholders involved in efforts to improve maternal health. Learning communities supported by other public-sector entities are also critical to expanding collaboration. HHS is aware of examples such as one group that fosters collaborative learning between Medicaid agencies and state universities to improve data capacity, and another that facilitates peer-to-peer learning and discussion among senior clinical leaders of state Medicaid programs.

A Call To Action

We can and must do better in making the United States one of the safest countries in the world for women to give birth. Research using California data indicates that as many as two-thirds of pregnancy-related deaths are preventable, and implementing evidence-based measures to reduce maternal mortality has been shown to cut mortality by as much as half, although reducing disparities has proven more challenging. The road to progress can be paved, in part, by improving our nation’s maternal health data infrastructure by collecting, sharing, and linking appropriately safeguarded, comprehensive data on the characteristics of individuals, the care they receive, and their non-clinical needs.

Read more on Health Affairs Blog