The U.S. birth certificate system, the nation’s primary tool for tracking severe maternal morbidity, significantly undercounts serious maternal morbidity complications and masks the true scale of racial disparities, according to a new nationwide study comparing birth certificate records with hospital claims data.

A team led by LDI Senior Fellow Beth Pineles, MD, PhD, an Assistant Professor of Obstetrics and Gynecology at the Perelman School of Medicine, provided new insight into the nation’s maternal mortality crisis, showing that many complications go unreported, especially among Black and Hispanic patients, potentially skewing efforts to measure and address worsening maternal outcomes.

“This undercount of maternal morbidity events in the federal Birth Certificate system underrepresents the amount of morbidity there is in the United States,” said Pineles. “In terms of measuring inequities, this means that if you use the birth certificate system to study inequities in maternal morbidity, you would be underrepresenting the amount of inequity compared to what was found in a national commercial all-payer database of billing and administrative records.”

Titled, “Differential Reporting of Severe Maternal Morbidity on US Birth Certificate and Claims Data by Race and Ethnicity,” the study is published in the May edition of the journal Epidemiology.

Clarification of “Birth Certificate”

When most people think of a “birth certificate,” they picture the small, one-page form new parents receive from the hospital. In practice, the term also refers to a much larger record of birth information collected by states and compiled into the Centers for Disease Control and Prevention’s National Vital Statistics System .

Although this larger federal system’s name suggests a focus on the infant, the formal U.S. birth certificate is a dual-capture document that collects standardized information about both the newborn and the mother. Drawing from the medical record and maternal worksheets completed at delivery, it includes demographic details such as race and ethnicity, age, and education, along with selected clinical risk factors and outcomes.

Manual Collection and Transfer

Pineles explained that in most areas, these data are manually collected by hospital personnel and sent on to their state’s vital records systems, which in turn, send them on to the federal National Vital Statistics System. This federal system serves as the nation’s only near-complete dataset covering all registered births. Researchers regularly use this data to track maternal mortality, monitor population-level trends, and inform public health policy.

However, this national system has important limitations. Much of its information is recorded at the hospital level through simplified checkboxes rather than detailed clinical coding. Data quality can vary depending on hospital staff documentation practices. For instance, some conditions are underreported or misclassified. The timing and severity of complications are often not well specified.

“There are a lot of opportunities for mistakes during manual input,” said Pineles.

As a result, the birth certificate system is a less sensitive instrument than hospital claims data, which are tied to billing codes and clinical documentation that include ICD diagnosis codes, procedures performed during surgery, medications, services billed, length of stay, and readmissions.

National Comparative Analysis

In the study, researchers analyzed more than 3.4 million birth certificate deliveries and 3.4 million deliveries in the Premier Healthcare Database, an all-payer database of billing and administrative records from hundreds of hospitals across the country. They compared incidence rates of blood transfusion, hysterectomy, intensive care unit admission, uterine rupture, and third- and fourth-degree perineal lacerations between 2019 deliveries in federal birth certificate records and the Premier Healthcare Database.

Blood transfusion rates serve as a key proxy for how well hospitals manage postpartum hemorrhage, the leading cause of maternal morbidity, with higher rates potentially signaling delays in recognizing or treating excessive bleeding. Unplanned hysterectomy during or immediately after childbirth is a rare but critical “sentinel event,” typically reflecting life-threatening complications such as uncontrolled hemorrhage or infection and indicating whether earlier interventions were effective. Rates of ICU admission capture the overall severity of maternal complications, with unplanned admissions suggesting failures to prevent or stabilize conditions such as preeclampsia, sepsis, or respiratory distress. Uterine rupture, a catastrophic tearing of the uterus often linked to prior cesarean scars, is a key indicator of the safety and management of vaginal birth after cesarean. Finally, third- and fourth-degree perineal lacerations, which involve severe tearing of the anal sphincter or rectum, reflect the quality of obstetric care, particularly around operative deliveries, and are closely tied to long-term maternal health outcomes.

Missing Data

“When we compared the two datasets, we found that consistently the birth certificate system was reporting fewer maternal morbidity events compared to the Premier dataset,” said Pineles. For example, for hysterectomy, the birth certificate system was reporting 46 per 1,000 hysterectomies, whereas the Premier dataset was reporting 148 hysterectomies per 1,000 patients. This was a gap of about a third, so the birth certificates were reporting about a third of the hysterectomies that the Premier dataset was reporting. And what was surprising was the size of this gap — this is a very large gap — two-thirds of the hysterectomies reported in the Premier dataset were not reported in the birth certificate.”

For severe lacerations, rates were 1,214 per 100,000 vaginal deliveries in birth certificates and 2,204 per 100,000 in claims data.

Masking Racial Disparities

In addition, for White patients, birth certificates captured about 50% of transfusions seen in claims data, but for Black and Hispanic patients, they captured about 29% to 39%.

Racial disparities appeared smaller or sometimes absent in birth certificate data. Compared to white patients, Black patients had only 16% higher odds of transfusion in birth certificate data, compared with 84% higher odds in claims data.

Overall, the study argues that U.S. birth certificates are not just incomplete but systematically biased in ways that:

• Underestimate the true burden of several major complications

• Understate racial disparities

• Potentially mislead research, policy, and public health surveillance

“One solution to this problem includes electronic capture of events from the maternal medical record into the birth certificate,” said Pineles. “This is actually not currently happening in the majority of jurisdictions. Instead, they have a birth registrar who may have been variably trained capturing events from the maternal medical record and manually transferring them into the birth certificate.”

“However,” she continued, the CDC’s National Vital Statistics System has begun working with states to collect information electronically from the electronic medical record into the birth certificate. And, moving forward, this initiative is very important for reducing errors and capturing all of the events that happen during a given birth.”

---

The study, “Differential Reporting of Severe Maternal Morbidity on US Birth Certificate and Claims Data by Race and Ethnicity,” was published in Epidemiology in May, 2026. Authors include Beth L. Pineles, Anthony D. Harris, Lisa Pineles, Esa M. Davis, KS Joseph, Enrique Schisterman, Laurence S. Magder, and Katherine E. Goodman.

---

---