Medicaid plays a vital role in maternal and child health, covering nearly half of the births in the United States—approximately 1.5 million annually. Despite this, clinical evidence and real-world evidence in these fields has traditionally centered on data from commercially insured populations, which largely reflect higher-income people with employer-sponsored insurance. As a result, lower-income families and pregnant individuals—key groups that could benefit the most from targeted health studies—are often overlooked, leaving critical gaps in understanding and care.
Although Medicaid administrative data encompass the full continuum from prenatal to postnatal care for both mothers and infants, they remain significantly underused, hampered by limited awareness, restricted access, and the complexities of linking maternal and infant records. For researchers in maternal and child health, getting started can be daunting. It’s often unclear how to access the data, where to begin, and what insights the data might reveal. Yet closing this gap is essential to unlocking the powerful longitudinal insights Medicaid data can offer into maternal and child health outcomes.
To help realize this potential, Mathematica drew on our access, expertise, and years of experience with Medicaid data to successfully link approximately nine million mother–infant pairs from 2016 to 2023, capturing nearly two-thirds of all Medicaid-covered births during that period.
The power of Medicaid data for real-world evidence on maternal and child health
Medicaid and Children's Health Insurance Program (CHIP) data from the Centers for Medicare & Medicaid Services (CMS) contain a wealth of information on enrollment, eligibility, and closed claims, covering inpatient stays, outpatient visits, prescriptions, vaccinations, and specialized services such as neonatal intensive care unit (NICU) services. Generally, women remain continuously enrolled in Medicaid during the perinatal period when they may be exposed to a variety of factors that can affect both the mother and child. Infants often remain enrolled through early childhood, providing reliable follow-up at key developmental milestones: 12 months, two years, and up to four years after birth. During the COVID-19 public health emergency (March 2020 through mid-2023), coverage for both mothers and infants was even more stable, as the federal government paused disenrollments.
In addition to covering a large share of births and enabling near-continuous follow-up during critical exposure and early developmental stages, CMS Medicaid data offer key advantages over commercial claims, such as the inclusion of low-income populations, detailed enrollment information, and provider identifiers. Unlike all-payer data sets that often reflect a limited subset of Medicaid managed care plans, CMS data include all enrollees nationwide, capturing fee-for-service and managed care encounters for a more complete view.
Once mother and child records are linked, researchers can address a wide range of critical questions, including the following:
- Quantifying in-utero exposures to therapies and maternal conditions
- Estimating the long-term burden of perinatal conditions
- Assessing the effects of maternal conditions and interventions on early childhood outcomes
- Evaluating long-term safety of in-utero exposures to therapies
- Quantifying healthcare utilization and costs associated with high-burden events, such as preterm births, NICU admissions, cesarean deliveries, and postpartum complications
How we linked mothers and children in Medicaid and CHIP
Mathematica is a long-time trusted partner to the Center for Medicaid and CHIP Services (CMCS), leading efforts to aggregate, standardize, and analyze data for the entire Medicaid and CHIP populations. In collaboration with CMCS, we produce the Transformed Medicaid Statistical Information System (T-MSIS) Analytic Files (TAF)—the largest, highest-quality, research-ready data set covering all Medicaid and CHIP enrollees across all 50 states, the District of Columbia, and U.S. territories that captures comprehensive information on enrollment, demographics, service utilization, payments, providers, and managed care plans.
Recently, we successfully linked approximately nine million mother–infant pairs in TAF Research Identifiable Files (RIF) data from 2016 to 2023—nearly two-thirds of all Medicaid births—using a scalable, privacy-preserving analytic framework. We started by identifying deliveries in national Medicaid data using diagnosis and procedure codes, resolving overlapping claims and inconsistent delivery dates to isolate distinct inpatient births. In parallel, we compiled infant records on birth, accounting for variation in Medicaid data elements and state-specific billing practices. Then, we applied a robust matching algorithm to maximize the match rate and accuracy and enable longitudinal tracking of maternal and infant outcomes across multiple pregnancies. We assessed the match patterns in aggregate against national birth certificate data from the National Vital Statistics System.
Looking ahead
As additional years of TAF data become available, opportunities for maternal–child health research will continue to grow. Medicaid data can be difficult for new users to navigate due to significant variation across states, evolving policy landscapes, and the complexity of its file structures.
By partnering with Mathematica, clients benefit from our access to comprehensive national Medicaid data, our deep experience improving and analyzing data, and a flexible modular analytic framework designed to adapt to varying client needs and research questions. Paired with a proven mother–infant linkage methodology, we tackled data standardization, validation, and infrastructure challenges from the outset. This enables clients to confidently launch maternal and child health research initiatives on day one without needing to build internal capacity. Together, we will work to transform data and analytics into smarter decisions, stronger policies, and better health outcomes.
To learn more about how Mathematica can help your organization conduct complex analyses of national Medicaid data, including linkage of mothers and infants, please contact us today.
