The Midwife Purge

At the turn of the twentieth century, approximately half of all births in the United States were attended by midwives, and the majority of those midwives — particularly across the rural South — were Black women working within community networks that had carried American maternity care through slavery, Reconstruction, and the Jim Crow consolidation that followed. By 1935, midwife-attended births had fallen below thirty percent. By 1950, below ten percent. By 1970, below one percent.

The replacement that produced these numbers was not a market choice. It was a coordinated thirty-year campaign of state licensure laws, public-health authority, and professional-society lobbying, executed primarily through the apparatus of the Sheppard-Towner Act of 1921 and the state-level Maternity and Infancy programs it funded — and it succeeded in eliminating the only category of maternity provider that had ever served Black women in the rural South at scale.

The clinical inheritance of this purge is visible in the maternal mortality data of 2026. Black women in the United States die in childbirth at approximately three times the rate of white women.¹ The disparity has been stable, in this approximate ratio, for the entire post-war period — a fact that the standard explanations (insurance, hospital quality, comorbidity prevalence) do not, in the end, fully account for. The granular epidemiology suggests that a significant fraction of the disparity is attributable to the breakdown of trust between Black patients and the medical-professional class that displaced their community-based providers a hundred years ago. The midwife purge is not the only cause of the disparity. It is, however, the institutional event that produced the structural conditions inside which the disparity has been continuously reproduced.

This essay is about how the midwife purge happened, who organized it, what statutory vehicles carried it, and what the present-day implications are for the Clinical Partner Network — Diosa Ara's federally distributed maternity-care infrastructure project — and for the renewed federal interest, twenty-five years into the twenty-first century, in restoring midwifery and doula access to the populations the 1900–1930 purge most directly harmed.

The Setup

The American obstetrical landscape in 1900 was, on close inspection, a deeply bifurcated system organized along lines of race, geography, and class. In urban centers — Boston, Philadelphia, New York, Chicago — the professionalization of medicine that the AMA had spent the second half of the nineteenth century pursuing had begun to consolidate hospital-based, physician-attended delivery as the standard of care for middle- and upper-class white women. The lying-in hospitals of these cities served as both clinical settings and as training grounds for the emerging obstetric specialty.² In the rural South, in immigrant urban neighborhoods, and across the Mississippi Delta, however, the operative reality was different. Approximately fifty percent of all U.S. births in 1900 occurred outside the physician-hospital system. The providers attending those births were midwives — predominantly women, predominantly working within community networks, and, in the rural South, predominantly Black.

The "granny midwives" of the South — the term, contemporaneous but increasingly contested in modern scholarship, described women who had typically been delivering babies for twenty or thirty years and who occupied a position of substantial community authority — were, in operational terms, the entire maternity-care system available to Black women in counties where Jim Crow segregation had foreclosed access to white-staffed hospitals and where the small number of Black physicians who had survived the post-Reconstruction medical-school closures could not, geographically, cover the rural population.³ These women carried obstetric knowledge transmitted through apprenticeship chains that extended, in some Mississippi and Alabama Black-belt counties, back to women who had been brought to the United States as enslaved healers in the eighteenth century. The knowledge base was empirical, oral, and locally calibrated to the populations served. The mortality outcomes — to the extent the data permits comparison, which is poorly, because the federal birth-registration system did not achieve national coverage until 1933 — were not measurably worse than the contemporaneous physician-attended outcomes, and on several specific metrics (puerperal sepsis rates, in particular, before the introduction of antiseptic obstetric technique in physician practice) appear to have been measurably better.

The second piece of the setup is the Flexner Report. In 1910, the Carnegie Foundation commissioned Abraham Flexner, a Johns Hopkins-trained educator, to survey American medical education. The report's conclusions were sweeping and consequential: most American medical schools, Flexner found, were operating at substandard educational quality and should be either reorganized to a hospital-based, four-year, scientifically grounded model or closed.⁴ The recommendations were largely adopted, with one demographically devastating consequence: of the seven historically Black medical schools operating in 1910, five closed within the decade. Only Howard University and Meharry Medical College survived. The pipeline that might have produced a Black physician class large enough to replace the granny midwives with professional Black obstetricians was, by 1925, foreclosed.

I need to make this personal, because it is. I am a physician. I practice competent, careful medicine; I take excellent care of my patients. And if I had been born a hundred years earlier, none of that would have mattered — because I am a Black woman, and the profession I trained in had already bolted its doors against me. The Flexner Report shut down five of the seven Black medical schools in the country. Medicine was being remade as white and male by design. So I would not have been the doctor in this story. I would have been the midwife — the skilled Black woman attending births in her community, doing the work I do now, until the state licensed her out of existence and called it progress. The purge did not just erase a profession. It erased the women who would have been me.

What this means for the purge that followed is that the elimination of the midwife was not, in the rural South, accompanied by the substitution of an alternative community-based or community-trusted provider. The hospital-based physician model that replaced midwifery in white urban America was, in Black rural America, simply not available. The purge produced an absence. The absence has not been filled in the hundred years since.

The Mechanism

The Sheppard-Towner Maternity and Infancy Protection Act, signed by President Harding in November 1921 and operative through 1929, was, on the level of stated intent, the first federal grant-in-aid program for maternal and infant health in American history. The Act appropriated approximately $1.25 million annually for state-level Maternity and Infancy programs, administered by the Children's Bureau within the Department of Labor, and was passed in significant part through the political agency of the women's suffrage movement — the Nineteenth Amendment having been ratified the year before, and the demonstrated electoral muscle of women voters being, in 1921, an asset that the Harding Republicans were eager to court.⁵ The Act funded prenatal and infant-health education programs across every state, established maternity-and-infancy clinics in approximately three thousand counties, and produced the first systematic federal data on U.S. maternal and infant mortality.

The operational reality of Sheppard-Towner, however, was more complex than the political framing suggested. The Children's Bureau, under the direction of Julia Lathrop and her successor Grace Abbott, did indeed use the program to extend health-education resources to populations that had previously been underserved. But the state-level public-health authorities that administered the funds — particularly in the South — used those same funds to underwrite a parallel campaign of midwife registration, licensure, and supervision that was, in practice, the principal mechanism by which midwifery was eliminated as a category of professional practice.⁶ The state laws followed a recognizable template. Midwives were required to register with the state Board of Health. Registration required passage of a state-administered examination conducted in English (a barrier for many immigrant midwives in urban areas and for many older Black midwives in the South whose formal literacy was limited under Jim Crow educational segregation). Registration further required participation in state-administered "training" classes, which were typically taught by white nurses and physicians who explicitly conceptualized the training as a transition program toward eliminating midwifery in favor of physician-attended hospital delivery.

The Mississippi program, often cited as the archetype, is instructive. Under the Mississippi State Board of Health midwife-supervision program initiated in 1921 and expanded through the Sheppard-Towner appropriations, every midwife operating in the state was required to attend monthly "midwife club" meetings administered by white county health nurses.⁷ The meetings combined elementary public-health instruction with explicit moral and racial framing. Midwives were prohibited from administering certain herbal preparations, restricted in their use of certain delivery techniques, and required to refer to physicians in circumstances that, in the operational reality of a Mississippi Black-belt county in 1925, often meant referring to a physician who would not see Black patients or who would see them only on Jim Crow-segregated terms. The number of registered midwives in Mississippi fell from approximately four thousand in 1921 to a few dozen by the 1970s. The fall was not the result of patient preference. It was the result of an attrition mechanism specifically engineered to produce it.

The parallel mechanism in the urban North was the licensure of nurse-midwifery as a credentialed professional category and the simultaneous exclusion of immigrant midwives from the credentialing pathway. The Frontier Nursing Service, founded in 1925 by Mary Breckinridge in rural Kentucky, and the Maternity Center Association of New York, which trained the first American nurse-midwives in 1932, represented the institutional vehicle through which midwifery was permitted to survive in a regulated, professionalized form.⁸ The pathway was not open to most existing midwives. It required prior registered-nurse credentialing, which required nursing-school admission, which required educational and financial prerequisites that systematically excluded Black women, immigrant women, and rural women — the populations from which the existing midwife workforce was drawn. The "professionalization" of midwifery produced a new category of certified nurse-midwife that, by 1950, was overwhelmingly white and urban-trained, and that by 1970 numbered only a few hundred providers nationally. The new workforce was vastly smaller than the workforce it nominally replaced.

The political coup de grâce came with the non-renewal of Sheppard-Towner in 1929. The American Medical Association, which had opposed Sheppard-Towner from the outset on the explicit grounds that it represented "state medicine" — a phrase the AMA would deploy against every subsequent federal health-policy initiative for the next forty years — successfully lobbied the incoming Hoover administration not to renew the program.⁹ The state-level midwife-supervision programs continued under reduced funding through the early 1930s and were absorbed, in 1935, into the maternal and child health provisions of Title V of the Social Security Act, which continued the licensure regime without funding the community-based midwifery infrastructure that the licensure regime was eliminating. The structural mechanism for the purge — federal funding for state-level licensure programs that systematically attrited the existing workforce — was, by 1935, fully institutionalized. The purge continued, by inertia, for the next thirty-five years.

The Aftermath

The most immediate and measurable aftermath of the midwife purge is in the maternity-care-desert geography of 2026. The United States currently has approximately 1,104 counties classified by the March of Dimes as maternity care deserts — counties with no hospital offering obstetric services, no birth center, and no obstetric provider.¹⁰ The geographic distribution of those deserts maps, with strikingly high fidelity, onto the rural South Black-belt counties where the granny midwife was the only available maternity provider in 1920. The hospital-and-physician system that replaced the midwives in those counties was either never built at the necessary density or, where built, was closed in the rural-hospital closure waves of the 1990s, 2010s, and the post-COVID 2020s. The population that depended on the midwife is the same population that depends, today, on driving two to three hours to reach an obstetric provider, or on not reaching one at all.

The clinical inheritance is the maternal mortality disparity itself. The CDC (National Center for Health Statistics, 2022) reports Black maternal mortality at approximately 49.5 deaths per 100,000 live births and white maternal mortality at approximately 19 deaths per 100,000 — a roughly 2.6-fold disparity that persists across income and educational strata.¹¹ The disparity is not, on the available evidence, primarily attributable to comorbidity prevalence, insurance status, or hospital quality at the population level — although each of those factors matters at the individual level. The single most parsimonious explanation that fits the full demographic stratification of the disparity is the breakdown of the trust relationship between Black patients and the medical-professional class that displaced their community-based providers in the 1900–1930 window. The patient who arrives at an obstetric encounter with a hundred-year inherited skepticism about whether her concerns will be heard is, statistically, less likely to receive the early-warning escalation that prevents the preventable death. The mechanism that the midwife model encoded — provider known to the community, accountable to the community, present continuously rather than episodically — was the mechanism the purge eliminated.

The renewed federal and philanthropic interest in midwifery and doula care that has accumulated over the past decade is, in this analytic frame, a partial and belated restoration. The Black Mamas Matter Alliance, established in 2013 in response to the persistent disparity, has played a central role in re-establishing community-based perinatal-care models.¹² The federal Maternal CARE Act and adjacent legislative vehicles — including the Black Maternal Health Momnibus — have introduced, with mixed success, the framework for Medicaid reimbursement of doula services. Fourteen states had implemented Medicaid doula coverage as of 2025. Several federal grant streams — most operationally through HRSA and through the CDC's Hear Her campaign — have supported community-perinatal-health pilots. The Clinical Partner Network, Diosa Ara's distributed maternity-care infrastructure project, is the direct present-day descendant of the midwifery model the purge eliminated. Its fundamental claim, examined inside its historical inheritance, is that the community-based provider relationship the United States dismantled between 1900 and 1930 was the operational mechanism by which preventable maternal deaths were, in fact, prevented — and that restoring that relationship is a higher-yield intervention than incremental improvements to the hospital-physician model that replaced it. The maternal mortality crisis of 2026 is also, in part, the unresolved aftermath of the purge of 1921.

The bridge to the Critical Infrastructure series runs most directly through Essay 4, The Workforce (see The Quiet Dismantling, Week 6 — Workforce Exodus), which describes the present-day obstetric workforce collapse — the ban-state exodus of OB-GYNs, the residency-program contraction, the L&D-unit closure cascade, the rural-hospital obstetric-service withdrawal — as a contemporary repetition of the structural condition the midwife purge produced in the 1920s.¹³ The mechanism in 2026 is different — administrative defunding, professional-liability exposure, ban-state legal risk — but the population most directly affected, and the geographic distribution of the harm, are continuous with the 1920s purge. The federal-administrative posture that produces obstetric-care deserts in the 2020s is the descendant, organizationally, of the federal-administrative posture that produced midwife extinction in the 1920s. The pattern is not new. The pattern's victims are the same.

The purge converted American obstetrics from a community-based, racially distributed practice into a hospital-based, racially stratified one — and its effects have never been fully reversed.

What You Can Do

Vote in the 2026 midterm with explicit attention to the Senate races and House appropriations committees whose membership determines whether Medicaid doula expansion, the Maternal CARE Act, and the HRSA community-perinatal-health appropriations survive the FY2027 cycle.

Press your state legislators on full Medicaid doula coverage and on the state-level midwife-licensure modernization that, in most states, still operates under the regulatory architecture the 1921–1929 purge produced.

Apply to — or support an applicant to — the Clinical Partner Network, because the CPN model is the operational restoration of the community-based perinatal-care infrastructure the purge eliminated, and because the federal-administrative environment of 2026 will not, on its own, restore what the federal-administrative environment of 1921 took apart.

This is the second essay. The Long War continues.