The Case for Roe v. Wade: What Fifty Years of Evidence Says About Women's Lives

Weeks after the Supreme Court erased the constitutional right to abortion in June 2022, a 28-year-old medical assistant named Amber Nicole Thurman arrived at a hospital south of Atlanta needing a routine procedure. A rare complication from medication abortion had left tissue in her uterus, and she needed a dilation and curettage, the same operation performed after miscarriages every day. Georgia's six-week abortion ban had been in force for barely a month. Her hospital did not act for about twenty hours while infection spread. She died during surgery, on August 19, 2022, leaving a six-year-old son. Months later, Georgia's own maternal mortality review committee, a panel of state-appointed experts, judged her death preventable, a finding first made public by ProPublica in September 2024.

Thurman's death is one case in the largest re-litigation of Roe v. Wade since 1973. This article argues one side of that case: that the constitutional protections Roe established produced measurable, documented benefits for women's health and bodily autonomy, and that Dobbs v. Jackson Women's Health Organization is now eroding them. But the argument only deserves a hearing if it survives the fact that cuts hardest against it. In the national statistics, maternal deaths have fallen, not risen, since Dobbs. What follows includes that number, the measurement problems wrapped around it, and the places where the predicted harms did, verifiably, arrive. What it does not attempt is to settle the moral question of when the government's interest in protecting potential life begins. No dataset answers that, and Roe itself never pretended to be an answer to it.

The hospital ward that vanished

Before Roe, abortion was not rare. It was dangerous.

After the Civil War, Congress passed the 1873 Comstock Act, criminalizing contraception and abortion together as obscenity, and states enacted their own versions, including Connecticut in 1879, as legal historian Reva Siegel recounts in the Yale Law Journal Forum. By roughly 1880, most states had made abortion a crime at every stage of pregnancy except to save the woman's life, a history the Roe majority itself traced in detail.

By the mid-twentieth century, historians estimate that between 200,000 and 1.2 million illegal abortions occurred each year in the United States; the true figure is unknowable precisely because the procedure operated in a shadow economy. What can be measured are the corpses the system certified. Federal statistics recorded a few hundred deaths per year attributed to induced abortion in the early and mid-1960s, and in 1965 abortion accounted for about 17 percent of all officially reported pregnancy-related deaths, according to Guttmacher's review of the pre-Roe record. Public hospitals in cities including Chicago and Los Angeles kept entire wards for septic abortion patients, women arriving hemorrhaging or infected after illegal procedures. Those wards closed in the years after legalization because the caseload collapsed.

The deaths were never evenly distributed. From 1972 to 1974, the chance of dying from an illegal abortion was about twelve times higher for nonwhite women than for white women, a disparity documented in federal analyses cited by Guttmacher. Money bought safer care: referral networks, overseas clinics, physicians willing to stretch the "therapeutic" exceptions that some states allowed. Criminalization did not eliminate abortion. It sorted abortion by safety, and it sorted lethally.

What legal abortion did to the death statistics

In 1972, the last full year before Roe, the Centers for Disease Control attributed 39 deaths to illegal abortion and 24 to legal abortion nationally, per Guttmacher's historical analysis. By 1976, deaths attributed to illegal abortion had fallen to low single digits. Nothing about the procedure changed. What changed was where it happened and who performed it. The same operation that had been improvised in motel rooms moved into licensed clinics with trained clinicians, sterile equipment, and follow-up care.

The safety record that followed became one of the strongest in outpatient medicine. CDC surveillance has put mortality from legal abortion below one death per 100,000 procedures for years. Epidemiologists Elizabeth Raymond and David Grimes calculated in 2012, in Obstetrics & Gynecology, that a woman was roughly fourteen times more likely to die from carrying a pregnancy to term than from a legal induced abortion, using the mortality figures of that era. The Roe majority had made a version of this point from the bench, noting that early legal abortions carried mortality rates "as low as or lower" than childbirth, and that the high death rates at illegal "abortion mills" strengthened rather than weakened public-health regulation of the legal procedure.

Two honesty checks matter here. First, the dramatic mid-century decline in overall U.S. maternal mortality, from around 700 deaths per 100,000 live births in the 1930s to near 20 by the early 1970s, was driven mainly by antibiotics, blood banking, and better obstetrics. Roe cannot claim that credit. Its measurable domain was the abortion-attributable slice, which had accounted for roughly one in six reported pregnancy-related deaths in 1965 and was reduced to nearly none within a decade. Second, the pattern repeats internationally today: the World Health Organization estimates that unsafe abortion causes between about 5 and 13 percent of maternal deaths globally, overwhelmingly in jurisdictions where abortion is restricted; a 2020 analysis in The Lancet Global Health found that abortion incidence is not lower in countries that prohibit it. Restrictions move abortion. They do not remove it.

What privacy built beyond the procedure

Roe did not emerge from nowhere. In Griswold v. Connecticut (1965), the Court struck down a contraception ban by locating privacy in the "penumbras" of specific Bill of Rights guarantees; in Eisenstadt v. Baird (1972), it extended that protection to unmarried individuals as well. Roe (1973) then held that this right of privacy, whether grounded in the Fourteenth Amendment's liberty or in the penumbras, "is broad enough to encompass a woman's decision whether or not to terminate her pregnancy." Nineteen years later, Planned Parenthood v. Casey (1992) discarded Roe's trimester framework, re-anchored the right more squarely in Fourteenth Amendment liberty, and replaced strict scrutiny with the "undue burden" test: states could regulate abortion unless a rule placed a "substantial obstacle" before a woman seeking one before viability. Journalist Linda Greenhouse has traced how this contraception-to-abortion-to-marriage lineage, from Griswold through Obergefell, functioned as a single doctrine of intimate decisional freedom, in the Connecticut History Review.

The autonomy claim was not only doctrinal. It had testable consequences. Between 2008 and 2010, researchers at UCSF's ANSIRH recruited roughly 1,000 women from thirty abortion facilities, comparing women who received abortions with "turnaways" who arrived just past gestational limits and were denied. The Turnaway Study's findings were stark: women denied abortions were more likely to report poor physical health and chronic pain for years afterward, and more likely to fall into poverty, with measurably worse credit and financial records, than women who received the procedure, as published in the American Journal of Public Health and synthesized in Diana Greene Foster's 2020 book. The study's authors reported two pregnancy-related deaths among participants, both in the group denied abortions. The sample was too small to treat those deaths as statistics. They function instead as an early, grim illustration of the comparison Roe's critics rarely engage: the question was never whether abortions would occur, but whether the alternative to legal abortion was childbirth under whatever conditions, or an illegal procedure under worse ones.

The warning the Court heard in advance

Before Dobbs was argued, the American medical profession put a prediction in writing. On September 20, 2021, the American College of Obstetricians and Gynecologists and the American Medical Association led an amicus brief joined by more than seventy medical organizations, telling the Court (in docket 19-1392) that abortion is safe, that it is far safer than compulsory pregnancy continuation, that banning it would increase maternal illness and death, and that the burden would land hardest on low-income patients and women of color. The brief pointed out that Mississippi, the state defending its fifteen-week ban, already ranked at or near the bottom on maternal, fetal, and infant outcomes.

On June 24, 2022, the day the decision came down, ACOG issued a statement calling the ruling "a direct blow to bodily autonomy, reproductive health, patient safety, and health equity in the United States," and warning that abortion bans would worsen the country's maternal mortality crisis. The American Medical Association said it was "deeply disturbed," and the American Academy of Pediatrics issued a parallel warning about adolescents. The three dissenters in Dobbs put the legal consequence in one sentence: under laws like Mississippi's, "from the very moment of fertilization, a woman has no rights to speak of."

This was, in short, a forecast entered into the record by the people who actually treat pregnant patients. Four years on, it can be graded against data, and grading it honestly means starting with the number the forecast's critics now cite most.

What the national numbers actually show after Dobbs

The National Center for Health Statistics tracks maternal deaths, defined as deaths during pregnancy or within 42 days of its end from causes related to or aggravated by pregnancy. Its annual series reads: 20.1 deaths per 100,000 live births in 2019; 23.8 in 2020; 32.9 in 2021; then 22.3 in 2022, the year Dobbs was decided; and about 18.6 in 2023, the first full post-Dobbs year, according to NCHS Vital Statistics Rapid Release reports issued for each of those years (the 2023 figures were finalized in 2025). The 2021 spike was driven overwhelmingly by COVID-19, which disproportionately killed pregnant women, and the subsequent decline is a reversion toward the pre-pandemic level.

So the blunt national prediction, that overturning Roe would visibly raise American maternal mortality, has not been borne out in the federal series. The number went down. Any honest version of the pro-Roe argument has to start by saying so.

It also has to explain why this national series is a poor instrument for detecting the harm ACOG described. The definition covers only deaths from obstetric causes within 42 days of the pregnancy's end. A homicide, suicide, or overdose at month eight is not in it, even though research in Obstetrics & Gynecology has found homicide to be among the leading causes of death for pregnant and postpartum women. A woman like Amber Thurman, dying of sepsis after a delayed procedure, is a single case in a national count built from a few hundred certificates a year, small enough that dozens of cases would barely flex the decimal. State-level rates are published in multi-year rolling windows because annual counts are so small. The series runs about two years behind. And a separate federal system, the maternal mortality review committees, which judged that 84 percent of pregnancy-related deaths in 36 states were preventable over 2017-2019, confirms by its existence that death certificates alone do not answer the question anyone actually cares about, which is whether care failed.

The right test of ACOG's warning is therefore not the national aggregate. It is the counterfactual inside the states that banned: what happened there, against what would have happened otherwise. One thread in the aggregate data deserves keeping in view as we turn to that question. Even as the national rate fell, the racial chasm barely moved. In 2023, NCHS recorded about 50 maternal deaths per 100,000 live births for Black women, against roughly 15 for white women.

Where the harms appear instead

Granular evidence exists, and part of it is now beyond dispute.

ProPublica's "Life of the Mother" investigation documented, case by case, women who died after state bans changed the medicine around them: Thurman and Candi Miller in Georgia, whose deaths state reviewers deemed preventable; Josseli Barnica, miscarrying at seventeen weeks in Houston in September 2021 under Texas's SB8, who waited some forty hours for intervention while the fetus still had a heartbeat, then returned septic and died days later, a case more than a dozen physicians reviewed for ProPublica and called preventable; Nevaeh Crain, eighteen, sent home from Texas emergency rooms with undiagnosed sepsis in the fall of 2023 before dying of organ failure. The series, collected at ProPublica's investigation hub, won the 2025 Pulitzer Prize for public service. In February 2025, ProPublica also published a statistical analysis of Texas hospital records documenting a sharp post-ban rise in fatal sepsis among patients hospitalized while miscarrying, a small absolute count that moved in exactly the direction clinicians predicted when emergency miscarriage care became legally fraught.

The most consequential epidemiology, though, concerns infants rather than mothers. In JAMA Pediatrics in June 2024, a Johns Hopkins team led by Suzanne Bell reported that Texas infant mortality rose 12.9 percent in 2022 while the rest of the country rose 1.8 percent; infant deaths attributed to congenital anomalies rose almost 23 percent in Texas while declining elsewhere, an estimated excess of about 216 infant deaths. In February 2025, companion analyses in JAMA and a companion JAMA journal extended the method to all post-Dobbs ban states: infant mortality ran about 6 percent above the expected trend, an estimated excess of roughly 480 infant deaths in 2022 and 2023, about four-fifths of them in Texas, with the companion analysis finding the increase disproportionately large, about 9 percent, among Black infants. The mechanism is not mysterious. Bans force the continuation of pregnancies diagnosed with fatal anomalies, and those births produce infant deaths that would not otherwise occur. These numbers have been contested, critics point to Texas data quirks and argue over baselines, but the finding's concentration in anomaly deaths is exactly the signature the causal story predicts.

The profession's warning also survives in a third form: what providers say is happening to care. In KFF's 2023 National Survey of OBGYNs, 68 percent of obstetrician-gynecologists said the Dobbs decision had worsened maternal mortality, and 64 percent said it had worsened doctors' ability to manage pregnancy-related emergencies. Texas's own supreme court inadvertently documented the legal fog behind those numbers in May 2024, when it decided Zurawski v. State of Texas: twenty women who said they were denied emergency care lost, but the justices still felt compelled to spend an opinion clarifying that the ban's exception permits "reasonable medical judgment" and does not require death to be imminent. Exceptions that require judicial clarification to function are not functioning.

The patchwork and who pays for it

Since Dobbs, legality has become a matter of geography. When the 2024 election results and subsequent rulings settled, a dozen states were enforcing near-total bans, and four (Florida, Georgia, South Carolina, Iowa) enforced bans around six weeks, before many women know they are pregnant. Courts in North Dakota and Wyoming struck down total bans late in 2024; Georgia's six-week ban was struck down in September 2024, then reinstated during appeal weeks later. Voters directly protected abortion rights in seven of the ten states voting on abortion measures that November, but Florida's measure took 57 percent of the vote and still lost, because Florida requires a 60 percent supermajority, while Nebraska voters chose the restrictive measure over the protective one. Missouri voters overturned their state's total ban via Amendment 3, and Missouri lawmakers have since placed a measure on the November 2026 ballot asking voters to reinstate it. Nevada's protection must pass a second vote in 2026 to take effect. The current, day-by-day map is best read on Guttmacher's interactive state tracker, because individual states keep moving with each court order.

Demand did not follow the law down. The Society of Family Planning's #WeCount project recorded about 1.03 million clinician-provided abortions in 2023, the highest count in over a decade, and totals remained above a million in 2024. Medication abortion accounted for 63 percent of U.S. abortions in 2023 per Guttmacher; by 2024, roughly one in five recorded abortions was dispatched by telehealth, about half under state "shield laws" protecting clinicians mailing pills into ban states. Guttmacher's Monthly Abortion Provision Study estimated that about 150,000 people crossed state lines for abortion care in 2023, nearly double the 2020 figure. The first interstate shield-law prosecutions, Texas's civil suit and Louisiana's indictment of a New York provider, are now working through the courts.

The burden of this adaptation is lopsided by design. A patient in Houston now reaches legal care by taking days off work and driving nine hours to Wichita, if she can. The states that banned already had the country's worst maternal-health infrastructure: even before Dobbs, the Commonwealth Fund documented that states moving to ban abortion recorded maternal mortality roughly 60 percent higher than states that protected access, a gap reflecting poverty, rurality, and insurance coverage more than abortion law itself, but one that bans now sit on top of. More than a third of U.S. counties are maternity care deserts, lacking obstetric clinicians or birthing facilities, and they overlap heavily with ban states. The medical pipeline is bending the same way: the Association of American Medical Colleges found OB-GYN residency applications in ban states fell about 6.7 percent in the 2023 cycle while edging up elsewhere, a pattern that persisted into the next cycle, and roughly four in ten accredited OB-GYN residency programs sit in states that restrict the procedure residents may need to learn. When residency follows training, and training follows legality, the desert widens.

The strongest case for Dobbs, answered

The argument for Dobbs, at its strongest, is not about mortality curves. The Dobbs majority, in 597 U.S. 215, held that the Constitution's text is silent on abortion, that a right to it is not "deeply rooted in this Nation's history and tradition," pointing to the near-universal state bans in place when the Fourteenth Amendment was ratified, and called Roe "egregiously wrong from the start." On this last point the majority was not alone across the ideological spectrum. John Hart Ely wrote in the Yale Law Journal in 1973 that Roe was "bad constitutional law," and Ruth Bader Ginsburg, in her 1992 Madison Lecture at NYU, argued that the Court had moved too far too fast, presenting a sweeping rationale instead of building incrementally and interrupting a reform movement that was already liberalizing state laws. The structural claim followed: nine unelected judges settled a contested moral question, and Dobbs returned it to voters, who indeed went on to settle it state by state. Defenders add the data point this article has already conceded: the predicted national mortality spike did not appear, and every ban carries a life-of-the-mother exception.

A fair answer concedes the doctrinal critique where it is earned and separates it from the empirical record. That Roe's reasoning was contested, even by its friends, is a criticism of judicial craftsmanship, not evidence that the regime it produced failed; the regime's output is in the death statistics, and it held for half a century. The historical argument, meanwhile, proves less than it seems: the same 1868 landscape that featured abortion bans also denied married women property rights and kept them off juries, because the Fourteenth Amendment's drafters did not imagine women as its chief beneficiaries. On exceptions, the documented record now includes state-appointed reviewers calling specific deaths preventable, and a state supreme court clarifying its own exception in writing, which is what non-function looks like. On the missing national spike, the correct reply is not to deny the number but to insist on the right instrument: counterfactual state-level analysis, which is where the infant-mortality excess appeared, concentrated in exactly the death category the mechanism predicts. And on democracy, the record cuts both ways: a 57 percent majority failed in Florida under a supermajority rule, and Missouri is being asked to vote again because its first answer was not the one its legislature wanted. "The voters will decide" means, in practice, that a woman's protections depend on which side of a state line she can afford to cross, and on whether her legislature accepts her vote the first time.

None of this touches the value judgment at the center. Roe itself said the state has a legitimate, eventually compelling, interest in potential life; the dissent between the two sides is over who holds the authority and the risk when that interest collides with the pregnant person in front of a doctor. The pro-Roe position ultimately rests on where the medical consequences land. They land on women, and, as the old death statistics showed and the new disparities data still show, on some women far more than others.

Does the privacy rationale still stand?

Partially, and defensively. The Dobbs majority took pains to say that "nothing in this opinion should be understood to cast doubt on precedents that do not concern abortion," and Justice Kavanaugh's concurrence emphasized that the decision does not threaten contraception or marriage rulings. Justice Thomas wrote for himself alone that the Court should "reconsider" Griswold, Lawrence, and Obergefell. As a matter of current doctrine, then, the Griswold line formally holds, four years on. But the constitutional theory that unified it, substantive due process protecting intimate choices, lost its enforceable core when Roe fell, and the protection of everything adjacent to abortion now depends on judicial self-restraint rather than any stable rule.

The autonomy argument did not die. It migrated. Since 2022, the fight has moved to state constitutions and statutes: Michigan's 2022 amendment, then Missouri's, Arizona's, and Montana's 2024 measures; Kansas's 2019 state-court precedent that continues to shield access inside a restrictive region. Federal battles have receded to the edges: the Court dismissed the mifepristone challenge for lack of standing in June 2024, FDA v. Alliance for Hippocratic Medicine, leaving the drug available under the FDA's telehealth framework while state attorneys general pursue the same fight in lower courts; the emergency-care conflict between federal EMTALA law and state bans was dismissed as improvidently granted in Moyle v. United States in June 2024, and the Justice Department abandoned the Idaho suit in March 2025, leaving no resolved national floor for emergency obstetric care. The privacy rationale survives the way its subject now lives: state by state, statute by statute, ballot by ballot.

The argument that survives the fact-check

Here, in the form that can be defended before a skeptical reader, is the case for what Roe did and what its loss is doing.

Roe's framework coincided with, and by the mechanism evidence supports, caused the near-elimination of a documented class of deaths and a documented class of injuries: deaths attributed to illegal abortion fell from hundreds per year to nearly none within a decade, septic wards closed, and legal abortion became statistically far safer than childbirth. That is a measurable, lasting benefit, and it is the strongest empirical line in the pro-Roe case. The protections also carried measurable autonomy benefits, established by the Turnaway Study's comparison of outcomes for women granted and denied the procedure.

The reversal of those protections is eroding the gains, but not where the loudest rhetoric promised. The national maternal-mortality aggregate has fallen since Dobbs and cannot, at least through the most recent finalized data, be recruited as proof of the prediction ACOG entered into the record. Where the evidence does show the predicted mechanism at work is concentrated and specific: review-committee-verified preventable deaths in ban states, an excess of infant deaths tracking closely with anomaly diagnoses, degraded emergency obstetric care by physicians' own account, widening care deserts, and a training pipeline bending away from the states that most need obstetricians. These harms are real, documented, and disproportionately borne by poor, rural, Southern, and Black women, which is also why they sit so poorly inside blunt national statistics built on small counts and narrow definitions.

What remains genuinely unknown is the long arc of the ban-state data. Three more years of finalized mortality data, better death attribution, and replication of the infant-mortality findings outside Texas will show whether the concentrated harms scale or attenuate. That is the honest state of the record in 2026, and it is a stronger argument than the sweeping prediction would have been, because it does not require the reader to accept anything that has not already been verified.

In the years after Roe, the septic wards emptied because the law stopped treating a medical crisis as a crime. The question Dobbs returned to fifty legislatures and a dozen court systems is ultimately the question Amber Thurman's twenty-hour wait poses: when a woman is dying of a treatable complication in a hospital that has everything needed to save her, except legal certainty about whether it may act, whether the law that answers is written for the patient or against her.

Sources and references

• Roe v. Wade, 410 U.S. 113 (1973), opinion via Justia
• Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), opinion via Justia
• Dobbs v. Jackson Women's Health Organization, 597 U.S. 215 (2022), opinion via Justia
• Griswold v. Connecticut, 381 U.S. 479 (1965), opinion via Justia
• Eisenstadt v. Baird, 405 U.S. 438 (1972), opinion via Justia
• FDA v. Alliance for Hippocratic Medicine, 602 U.S. 367 (2024), opinion via Justia
• Moyle v. United States, 2024, docket and opinion via Oyez
• Zurawski v. State of Texas, Texas Supreme Court, May 31, 2024
• American College of Obstetricians and Gynecologists et al., amicus brief and coalition filing in Dobbs, September 2021, Supreme Court docket 19-1392
• ACOG, statement on the decision in Dobbs v. Jackson, June 24, 2022
• American Medical Association, statement on the Dobbs decision, June 24, 2022; American Academy of Pediatrics, statement, June 2022
• National Center for Health Statistics, Donna L. Hoyert, "Maternal Mortality Rates in the United States" (annual Vital Statistics Rapid Release reports for 2019 through 2023); NCHS maternal mortality portal
• CDC, "Pregnancy-Related Deaths: Data from Maternal Mortality Review Committees in 36 US States, 2017-2019," MMWR, 2022
• Rachel Benson Gold, "Lessons From Before Roe: Will Past Be Prologue?", Guttmacher Policy Review 6(1), 2003 (including Cates and Rochat's federal analyses of 1965-1974 abortion mortality)
• Elizabeth G. Raymond and David A. Grimes, "The Comparative Safety of Legal Induced Abortion and Childbirth in the United States," Obstetrics & Gynecology 119(2), 2012
• World Health Organization, "Preventing unsafe abortion," fact sheet
• Bearak J. et al., "Unintended pregnancy and abortion by income, region, and the legal status of abortion," The Lancet Global Health, 2020
• Reva B. Siegel, "How Conflict Entrenched the Right to Privacy," Yale Law Journal Forum, 2015
• Linda Greenhouse, "From Griswold to Roe and Beyond," Connecticut History Review, 2025
• John Hart Ely, "The Wages of Crying Wolf: A Comment on Roe v. Wade," Yale Law Journal 82, 1973
• Ruth Bader Ginsburg, "Speaking in a Judicial Voice" (Madison Lecture), NYU Law Review 67, 1992
• Diana Greene Foster et al., "Socioeconomic Outcomes of Women Who Receive and Women Who Are Denied Wanted Abortions in the United States," American Journal of Public Health 108(3), 2018; Diana Greene Foster, The Turnaway Study, Scribner, 2020
• Kavitha Surana, "Abortion Bans Have Delayed Emergency Medical Care. In Georgia, Experts Say This Mother's Death Was Preventable," ProPublica, September 16, 2024
• ProPublica, the "Life of the Mother" investigation, 2024-2025 (including reporting on Candi Miller, Josseli Barnica, Nevaeh Crain, and the Texas hospital sepsis analysis of February 2025); awarded the 2025 Pulitzer Prize for Public Service
• Suzanne O. Bell, Alison Gemmill, et al., "Infant Mortality After Texas' 2021 Ban on Abortion in Early Gestation," JAMA Pediatrics, June 24, 2024
• Alison Gemmill, et al., "Estimated Impacts of State Abortion Bans on Infant Mortality in the United States," JAMA, February 13, 2025, and companion racial-disparity analysis in a JAMA journal the same date
• KFF, 2023 National Survey of OBGYNs on the effects of Dobbs on maternal mortality and emergency care, June 2023
• Commonwealth Fund, "The U.S. Maternal Health Divide," December 2022
• Maeve E. Wallace et al., "Homicide During Pregnancy and the Postpartum Period in the United States, 2018-2019," Obstetrics & Gynecology, 2021
• Guttmacher Institute, "Interactive Map: U.S. Abortion Policies and Access After Roe"; Guttmacher Institute, Monthly Abortion Provision Study (travel estimates); Guttmacher Institute, medication abortion share of U.S. abortions, 2024
• Society of Family Planning, #WeCount reports on clinician-provided abortion totals and telehealth share, 2024-2025
• Association of American Medical Colleges, analyses of residency-applicant trends in states with abortion bans, 2023 and 2024
• March of Dimes, "Nowhere to Go: Maternity Care Deserts in the U.S."