Maternal health data: a roadmap for community advocates

A headline about maternal mortality is not yet an advocacy strategy. The useful work begins with asking what was counted, whose experience was captured, and whether the data can support action at the local level.

The stakes are not abstract. In 2021, the U.S. maternal mortality ratio was 32.9 deaths per 100,000 live births. For non-Hispanic Black individuals, it was 69.9 deaths per 100,000 live births. Pregnancy-related mortality rates among Black women have remained more than three times those among White women, including when income and education are taken into account. That pattern does not point to a single individual risk factor. It points toward systems: clinical bias, unequal treatment, fragmented care, chronic stress, insurance gaps, transportation barriers, and the social conditions in which pregnancy takes place.

For organizations working on reproductive health equity, maternal health data access for community advocacy is therefore not a technical side issue. It is part of the power structure. Data determines which deaths are investigated, which disparities become visible, and which solutions receive funding.

Start with the definition, not the headline

The first mistake in maternal health advocacy is treating every pregnancy-related death as the same kind of event. Surveillance systems use several distinct categories, and each answers a different question.

A maternal death is generally defined as the death of a person while pregnant or within 42 days of the end of a pregnancy from a cause related to or aggravated by the pregnancy or its management. Accidental or incidental causes are excluded from that definition. The 42-day window matters because it creates a specific boundary for a particular surveillance measure; it does not mean that health risks disappear after six weeks.

Other terms may include deaths that occur during pregnancy or after pregnancy but are not caused by the pregnancy, as well as deaths occurring later in the postpartum period. Those categories can be essential for understanding the full burden of pregnancy-associated mortality, but they should not be casually merged with maternal mortality or pregnancy-related mortality.

Before requesting local maternal mortality data or comparing two jurisdictions, write down the exact outcome being measured:

• Maternal mortality concerns deaths related to pregnancy or its management within the defined period.
• Pregnancy-associated mortality can include deaths occurring during pregnancy or within a specified period afterward, regardless of cause.
• Pregnancy-related mortality is a narrower determination that examines whether the death resulted from or was aggravated by the pregnancy.
• Maternal mortality ratio is typically expressed as deaths per 100,000 live births, which makes the denominator part of the meaning.
• Maternal mortality rate may use a different denominator, such as the number of people of reproductive age or person-years, depending on the system.

These are not interchangeable labels. A report can be accurate while still being unsuitable for the comparison an advocate wants to make. A statewide maternal mortality ratio may not tell you how many people died in a particular county. A pregnancy-associated count may capture deaths that a pregnancy-related review would classify differently. A small local number may be suppressed for privacy or too unstable to support a precise rate.

A number becomes useful for advocacy only when you know exactly what it counts—and what it leaves outside the frame.

Why the surveillance window changes the story

Pregnancy-related harm often unfolds across a longer timeline than the formal maternal mortality definition. A person may lose insurance after delivery, struggle to obtain treatment for postpartum depression, experience a delayed diagnosis of cardiomyopathy, or face a preventable emergency after leaving the hospital. If the advocacy question is about postpartum care, the six-week boundary may not capture the full pathway to harm.

That does not make the formal definition wrong. It means the definition must match the policy question.

If the goal is to evaluate obstetric emergency response, hospital delivery data and severe maternal morbidity measures may be more informative than mortality alone. If the goal is to understand postpartum mental health, a review limited to deaths within 42 days may miss relevant cases. If the goal is to examine racial disparities, the dataset should be reviewed for race and ethnicity classification, missingness, sample size, and the way multiple-race identities are recorded.

A responsible report states these limits plainly. It avoids turning a clean-looking chart into a claim the underlying data cannot support.

MMRIA gives advocates a common language

The Maternal Mortality Review Information Application, or MMRIA, is a CDC data system designed to support the work of multidisciplinary Maternal Mortality Review Committees. Its purpose is not simply to store death counts. It gives participating states and territories a common data language for reviewing cases, identifying contributing factors, and documenting opportunities for prevention.

That common language matters because maternal deaths rarely fit inside one institution’s records. A review may draw on vital records, medical charts, emergency department notes, prenatal records, social service information, interviews with family members, and other available documentation. The review committee then considers not only the immediate medical cause but also the circumstances surrounding the death and whether it might have been prevented.

For community advocates, MMRIA is best understood as a framework rather than a public search box. The existence of a standardized system does not mean that every local case file is openly downloadable. Detailed records may be protected by privacy rules, state review policies, data-use agreements, or other restrictions. The level of public access differs by jurisdiction.

Still, MMRIA-informed findings can shape a practical advocacy agenda. They can help organizations ask:

• Which causes of death appear repeatedly in state review findings?
• Were people able to access prenatal, delivery, and postpartum care?
• Did insurance coverage or a change in eligibility interrupt care?
• Were warning signs recognized and escalated?
• Did transportation, housing, food access, language, or immigration-related concerns affect the ability to seek treatment?
• Were there communication failures between hospitals, clinics, emergency services, and patients?
• Did a person’s race, gender identity, disability, or sexual orientation affect the care they received or the credibility given to their symptoms?
• Which recommendations have been made by the review committee, and who is responsible for implementing them?

The value of a review is not limited to identifying a clinical cause of death. It can expose the chain of decisions and constraints that made a bad outcome more likely.

Build a local data map before making a request

Community groups often lose time by asking a health department for “all maternal mortality data” without specifying the decision the information is meant to inform. A narrower request is easier to process and more likely to produce a usable answer.

Begin with a one-page data map. It should identify:

1. The geographic area. County, city, hospital service area, tribal community, or multi-county region may produce different results.

2. The time period. State review data may cover a multi-year window, while hospital quality data may be reported annually.

3. The population. Define whether the focus is on all births, Medicaid births, adolescents, Black women, Indigenous people, LGBTQ patients, rural residents, or another group.

4. The outcome. Mortality, pregnancy-related mortality, severe maternal morbidity, postpartum follow-up, prenatal care initiation, emergency transfers, or another measure.

5. The intended use. A budget request, public hearing, hospital meeting, grant application, community needs assessment, or media investigation may require different levels of detail.

6. The minimum necessary detail. Ask for aggregated information when individual-level records are not needed.

This approach separates a legitimate need for evidence from an overly broad request that may trigger privacy concerns or administrative delay.

Potential data holders can include state health departments, local public health agencies, Maternal Mortality Review Committees, Perinatal Quality Collaboratives, hospital systems, Medicaid agencies, vital statistics offices, and community health centers. Each may hold a different part of the picture. No single dataset is likely to explain the entire pathway from pregnancy to outcome.

Read disparity data without turning inequality into an individual diagnosis

The racial gap in maternal outcomes is one of the clearest reasons advocates need better data. But the way the gap is described can either clarify the problem or quietly misdiagnose it.

Black women in the United States experience pregnancy-related mortality rates more than three times higher than White women. The disparity persists across income and education levels. That does not mean income, education, housing, or employment are irrelevant. It means that socioeconomic status alone does not explain the disparity.

A strong maternal health disparities report treats race as a marker of exposure to unequal systems and treatment, not as a biological explanation. The analysis should examine how race interacts with:

• The quality and continuity of prenatal and postpartum care.
• Insurance type, coverage interruptions, and reimbursement structures.
• Distance to obstetric services and availability of emergency transport.
• Hospital quality and the concentration of high-risk deliveries.
• Clinical communication and whether patients’ reports of symptoms are believed.
• Language access and the use of qualified interpreters.
• Housing instability, food insecurity, environmental exposures, and workplace conditions.
• Experiences of racism, discrimination, and chronic stress.
• Immigration-related fears or documentation barriers.
• Disability access and the treatment of disabled patients.
• Gender-affirming and culturally responsive care for LGBTQ patients.

The data will not always measure each of these factors directly. That is a reason to combine sources, not a reason to pretend the factors do not exist.

Use denominators carefully

A count of five deaths and a rate of 50 deaths per 100,000 live births may describe the same small population, but they communicate different things. Counts show the number of families affected. Rates help compare populations of different sizes. Neither should be used without context.

Small-area rates can swing sharply from year to year because the number of births or deaths is limited. A multi-year average can make a pattern more stable, though it may also conceal a recent change. Suppressed data may protect privacy but make a local disparity harder to see. The solution is not to fill gaps with assumptions. It is to explain the limits and, where appropriate, use several years or combine quantitative data with community testimony and qualitative research.

When comparing racial groups, also inspect how race and ethnicity were classified. A dataset with substantial missing information may understate disparities. A dataset that collapses distinct communities into a single category may conceal differences among them. Indigenous communities, Pacific Islander communities, immigrants, and people who identify with more than one race can disappear inside broad categories that look tidy in a table.

A useful analysis asks three questions of every comparison:

• Are the groups defined consistently?
• Are the denominators comparable?
• Could the observed difference reflect data quality or access to reporting, rather than the full underlying disparity?

These questions do not weaken an advocacy claim. They make it harder to dismiss.

Turn preventability findings into a policy argument

Maternal Mortality Review Committees have found that more than 80%—in some summaries, up to 87%—of pregnancy-related deaths were preventable. The central implication is not that every death could have been avoided through perfect medical care. It is that many deaths involved an opportunity for intervention.

That distinction should guide how advocates use the statistic. Preventability is not a moral judgment about the person who died or a simple accusation against one clinician. It is a way to examine the points where a health system could have acted differently.

Those points may appear before pregnancy, during clinical care, or after discharge:

• A patient could not obtain timely prenatal care.
• A warning sign was documented but not escalated.
• A hospital lacked a reliable protocol for hemorrhage, hypertension, sepsis, or cardiac complications.
• A referral was made without transportation, insurance approval, or a confirmed appointment.
• Discharge instructions were not accessible in the patient’s language.
• A postpartum symptom was treated as routine recovery.
• A primary care or behavioral health connection was missing after delivery.
• A medication, appointment, or follow-up plan was disrupted by loss of coverage.
• A patient’s concerns were minimized or not incorporated into the care plan.

The policy question becomes specific: which preventable factors are recurring, and what authority or funding is required to address them?

Connect each finding to an accountable actor

Data has more force when every recommendation names a responsible institution and a mechanism for follow-through. For example:

This structure prevents a familiar failure mode: a report that recommends “more awareness” but assigns no duty to anyone. Awareness may be useful, but it is rarely a measurable intervention by itself.

Perinatal Quality Collaboratives can help translate findings into quality-improvement work. HHS Perinatal Improvement Collaborative hospitals have used more than 150 clinical and non-clinical measures. The significance of that breadth is practical: maternal health quality is not only an operating-room issue. It also involves referral systems, communication, patient experience, discharge planning, workforce capacity, and social needs.

Community advocates should ask to see not only whether a hospital participates in a quality initiative, but what it measures, how results are stratified, whether patients help define the measures, and whether the findings lead to changes that can be tracked.

The strongest maternal health demand is not “release the data.” It is “show the data, name the gap, assign responsibility, and publish the follow-through.”

Build a public health department data transparency strategy

Public health department data transparency is not the same as releasing every record. Privacy protections are essential, particularly in small communities where a combination of age, race, facility, and outcome could identify a family. Transparency means making the measurement system understandable and the decision process visible.

A community organization can ask a health department to publish or explain:

• Definitions for maternal, pregnancy-associated, and pregnancy-related deaths.
• The years included in each report.
• The geographic level at which data is reliable.
• Rules for suppressing small numbers.
• Race and ethnicity categories, including how missing data is handled.
• Whether the findings come from vital records, MMRC review, hospital reporting, or another source.
• How preventability and contributing factors are determined.
• Which recommendations remain open and when progress will be reviewed.
• Whether community members, patients, and families participate in the review or implementation process.

The request should also distinguish between data release and data interpretation. A dashboard may display a rate without explaining a change in the denominator. A press release may cite a statewide trend that says little about a rural county. A hospital may report improvement in one measure while patients continue to experience barriers not captured by the metric.

A transparent system explains not only the result but the method.

Make data requests specific and usable

When contacting a public agency, use plain language and specify the format that would help. Ask whether the information exists as a public report, an aggregate table, a dashboard, or a record that requires a formal request. If the agency cannot provide the requested data, ask what comparable measure is available and why the original request cannot be fulfilled.

A productive request might seek multi-year, aggregated counts or rates by county and race, accompanied by the data dictionary and suppression rules. Another might ask for the status of recommendations issued by an MMRC, including the agencies assigned to implement them. A hospital-focused request might seek de-identified, aggregate measures of severe maternal morbidity, postpartum follow-up, transfer patterns, or patient-reported experience.

Avoid requesting more detail than the advocacy question requires. Individual medical records are rarely necessary for a public campaign and can create privacy risks. The goal is evidence that can be responsibly used, not the largest possible spreadsheet.

Work with community evidence when the dataset is incomplete

Formal surveillance is indispensable, but it is not a complete account of reproductive healthcare. People who never reached a prenatal clinic, moved between states, lacked stable housing, or avoided institutions because of discrimination may be poorly represented in administrative records. Patient testimony and community-based research can identify barriers that official data does not measure.

That evidence should not be treated as anecdotal decoration. It can be organized systematically through listening sessions, confidential interviews, surveys, patient advisory boards, and partnerships with doulas, midwives, community health workers, disability organizations, LGBTQ health groups, and local reproductive justice organizations.

The safeguards matter:

• Do not collect names or identifying details unless there is a clear, secure reason.
• Explain how stories will be used and whether participants can withdraw.
• Pay community members for their time when possible.
• Avoid asking people to repeatedly retell traumatic experiences without support.
• Separate a person’s story from any claim that cannot be established by the available evidence.
• Return findings to the community before publishing them widely.
• Let participants challenge the interpretation.

Community evidence can reveal that an official measure is poorly designed. For example, a clinic may report high postpartum visit completion while patients describe visits that were difficult to schedule, inaccessible by transit, conducted without interpretation, or unable to address urgent symptoms. Both facts can be true. The tension between them is itself a finding.

Use federal initiatives as leverage, not as a substitute for local accountability

Federal maternal health initiatives can provide language, funding priorities, and policy openings for local advocacy. The White House Blueprint for Addressing the Maternal Health Crisis, the Kira Johnson Act, the Social Determinants for Moms Act, and the Perinatal Workforce Act represent different approaches to maternal health equity, including attention to care systems, social conditions, and workforce capacity.

Their usefulness to a community organization depends on status, funding, implementation, and jurisdiction. A proposal is not the same as an enacted program. An announced initiative is not the same as money reaching a local clinic. A state plan is not the same as an enforceable service.

Use federal policy documents to ask concrete local questions:

• Which recommendations has the state adopted?
• What funding is available to community health centers, doulas, midwives, or public health programs?
• Are workforce investments reaching communities with the highest burden?
• Are social needs being addressed through durable programs or short-term pilots?
• How will outcomes be measured, and will results be publicly reported?
• Are patients and community organizations involved in oversight?
• What happens when a program ends or funding changes?

The legislative pathway also changes the advocacy strategy. A budget request, regulatory comment, public hearing, agency meeting, and hospital board presentation each require a different kind of evidence. Maternal health data becomes more effective when it is paired with a specific decision point: preserve a program, expand a service, change a protocol, fund a workforce, publish a measure, or investigate a disparity.

A practical workflow for using health statistics for advocacy

A community group does not need a large research department to begin. It does need a disciplined process that keeps the data connected to lived experience and an achievable demand.

1. Define the decision you want to change

Do not begin with the broad goal of improving maternal health. Identify the decision-maker and the decision. You may be seeking postpartum transportation funding, a hospital hemorrhage protocol, a public report on racial disparities, continuous insurance coverage, or investment in community-based providers.

2. Choose the smallest set of measures that can answer the question

A campaign about postpartum care may need follow-up rates, coverage continuity, emergency visits, and patient-reported barriers. A campaign about hospital safety may need severe maternal morbidity, emergency response measures, transfer patterns, and review recommendations. More indicators do not automatically produce a stronger case.

3. Establish the definitions and time frame

Record the surveillance definition, denominator, geographic scope, race and ethnicity categories, and years covered. If two sources use different definitions, say so rather than placing the numbers in a single comparison.

4. Look for a pattern across more than one source

Compare state reports with hospital quality information, community health center experience, Medicaid data, or patient testimony when appropriate. A repeated pattern across independent sources is more persuasive than a dramatic figure with unclear provenance.

5. Check for missing voices

Ask who is absent from the dataset: people without stable care, rural residents, incarcerated people, undocumented patients, disabled patients, LGBTQ patients, people who experienced pregnancy loss, or those who died outside a hospital. The absence may reflect a measurement problem rather than an absence of risk.

6. Translate the finding into a measurable demand

Replace a general request for action with a defined proposal. Specify the institution responsible, the resources required, the implementation date if known, and the measure that will show whether the change occurred.

7. Publish the method alongside the message

A short methods note can explain definitions, years, sources, suppression, and limitations. This protects the campaign from avoidable criticism and allows other organizations to reproduce or challenge the analysis.

What responsible maternal health reporting sounds like

Language shapes policy. A report should avoid describing racial disparities as the result of supposedly risky behavior by the people most affected. It should avoid presenting preventability as proof that every death resulted from one clinician’s mistake. It should not imply that a single intervention can overcome the effects of hospital closures, insurance instability, racism, inadequate transportation, and under-resourced public health systems.

More precise language is also more forceful. Instead of saying that a community is “high risk,” describe the conditions that create risk. Instead of saying that patients are “noncompliant,” ask whether appointments, transportation, interpretation, medication access, and respectful communication were available. Instead of treating a disparity as a fixed characteristic of a population, examine which institutions produce and can reduce it.

The objective is not to soften the findings. It is to place responsibility where the evidence points.

The route from data to change

Maternal health data access for community advocacy is a route, not a destination. The work moves from definitions to sources, from sources to patterns, and from patterns to accountable decisions. MMRIA and Maternal Mortality Review Committees provide an important structure for understanding preventable deaths. Perinatal Quality Collaboratives can help turn recurring findings into clinical and operational improvements. Public health departments can make the system more trustworthy by explaining what they measure and how recommendations are tracked. Community organizations can close the gaps by documenting experiences that administrative datasets miss.

The most credible advocacy does not claim that data can speak for itself. Data requires interpretation, and interpretation requires proximity to the people living with the consequences. When advocates pair rigorous definitions with patient knowledge, racial equity analysis, and a specific policy demand, statistics stop being background information. They become a way to identify responsibility—and to insist that preventable harm is not accepted as normal.