Fujimori's reproductive-health program and forced sterilization

A program promising reproductive choice made refusal unsafe

Peru approved the National Reproductive Health and Family Planning Program (PNSRPF) on February 6, 1996, placing its execution, coordination, supervision, and evaluation under the Ministry of Health. The program promised wider access to contraception and the freedom to decide whether and when to have children in a country where access had been sharply unequal.¹ From 1996 through 2000, however, the government of Alberto Fujimori also made permanent surgical contraception a visible measure of program achievement. Central pressure traveled down the health hierarchy, and many poor, rural, campesino, and Indigenous people encountered “choice” through professionals who held medical authority and controlled information.²

Sterilization is not itself the harm. When freely chosen with intelligible information, time, alternatives, clinical safety, and the continuing right to refuse, it can be valuable care. The institutional failure was to organize an irreversible decision as campaign throughput. A signature could be recorded as consent even when the person did not understand the form, was deceived, faced repeated pressure, or learned what had happened only after surgery. The program could report a completed service while excluding the quality that made the service legitimate: the patient's autonomous decision. The Inter-American Court defines valid consent for sterilization as prior, free, full, and informed; it also holds that pressure or coercion vitiates consent and can make the procedure forced sterilization.³

A neglected service became a presidential performance priority

The program joined public-health language to a development argument that associated high fertility with poverty. Expanding reproductive care answered a real need, especially beyond cities. But when poor and Indigenous women's fertility was treated as a cause of national underdevelopment, the state could regard fewer births as an outcome achieved on the population rather than a choice protected for each person.¹

The program was broader than sterilization. It also offered reversible contraception and maternal and child-health services. A peer-reviewed quasi-experimental study estimated that earlier program exposure increased use of both temporary and permanent contraception and reduced infant mortality, partly through longer breastfeeding. Those aggregate benefits were unequal: the authors found few positive child-health effects for Indigenous children whose mothers had been sterilized, and their data could not distinguish voluntary from coerced operations.⁴ Service expansion and coercion therefore belong in the same history without one cancelling the other.

The Fujimori government elevated family planning, and the Ministry of Health translated policy into campaigns, reporting expectations, and numerical goals. Regional directorates and facilities received pressure to produce results. Staff organized surgical campaigns, recruited candidates, scheduled teams, and returned completed procedures through the reporting chain. Resources and attention could favor surgical contraception because it was countable, permanent, and legible to senior officials. The Court's factual record describes monthly expectations, performance comparisons, incentives, and campaigns that favored surgical contraception. It also records Peru's position that the figures were planning estimates rather than mandatory quotas.²

A later Ministry of Health commission reported 215,227 tubal ligations and 16,547 vasectomies from 1996 through 2000. It said that 56 official documents showed pressure and direction running from high levels through ministries and regions.⁵ Those totals describe procedures, not a proven count of forced sterilizations. A 2002 Ombudsman's report separately analyzed 773 complaints received from June 1999 through September 2002; its categories and findings included 61 complaints of surgical contraception without consent and 346 involving failure to observe a reflection period.⁶ Procedure totals, administrative complaints, entries in the survivor registry, and case-specific legal findings answer different questions. None supplies a case-by-case count of every coerced operation.

The surviving public evidence is also gendered. Official totals include men who underwent vasectomy, but the accessible case law, survivor organizations, and ethnography assembled here chiefly preserve women's experiences. They do not establish how many men were coerced, how coercion differed by gender, or how the program affected people outside binary administrative categories.⁵

Targets reorganized professional judgment at the clinic door

A numerical expectation becomes dangerous when frontline workers cannot report that informed refusal is a successful outcome. Under target pressure, a clinician who protects a patient's hesitation can appear unproductive; a manager whose region reports fewer procedures can appear resistant. Staff may intensify home visits, minimize risk, withhold alternatives, or treat a signature as a compliance step. A patient then faces not one neutral medical offer but a coordinated institution returning until the desired answer is obtained. Official investigations documented numerical targets, incentives, missing counseling, and failures to honor a 72-hour reflection period; they do not establish that every worker acted from the same motive or that every operation was coerced.²⁶

Language and status made that encounter unequal. Information and consent forms could be in Spanish for Quechua-speaking or other Indigenous patients. Medical professionals could invoke expertise that rural patients had little practical power to challenge. Some survivors describe being approached during childbirth, while seeking care for a child, or shortly before an operation—conditions incompatible with unhurried deliberation. Others report threats, deception, insulting descriptions of their fertility, or promises tied to household welfare.⁷⁸

The Association of Peruvian Women Affected by Forced Sterilizations preserves testimony from survivors including Rute Zúñiga, Esperanza Huayama, and Josefina Quispe. Their accounts describe force, deception, procedures they did not understand, pain, abandonment by partners, and humiliation inside their communities. These are not merely illustrations of a policy record. Survivors identify what the record omitted: how professional and gender hierarchy entered the body, and how harm continued after a facility counted the operation as complete.⁷

An ethnographic study conducted in Cajamarca in 2016–2017 adds a consequence that legal and fertility-centered accounts often flatten. Eighteen women described debilidad, a loss of strength connecting bodily pain to agricultural work, household labor, income, marital conflict, emotion, and community standing. The study does not establish clinical causation or represent every survivor; it shows why repair cannot be reduced to infertility or a completed surgical record when affected women describe a wider damaged lifeworld.⁹

Administrative segmentation concealed the whole act

The organizational chain distributed responsibility. National leaders could claim a voluntary rights-based program. Regional officials could transmit targets without meeting the people affected. Facility managers could organize campaigns. Community workers could recruit. Clinicians could rely on a signed form. Data staff could aggregate completed procedures. If complications appeared later, another facility or family might absorb them. Each role could point to another layer while the patient experienced one state.²⁶

The information system privileged completion over consent quality. It could count an operation but not whether the explanation was understood in the patient's language, whether reversible methods were genuinely available, whether refusal changed access to care, whether staff had visited repeatedly, or whether a person experienced chronic pain and economic loss afterward. Adverse outcomes and complaints traveled more slowly than achievement data and lacked equal executive force.⁶⁹

Celia Edith Ramos Durand shows the lethal stakes of that separation. Health workers repeatedly visited her before a July 1997 tubal ligation at a low-level clinic where 15 women were operated on that day. The clinic was not equipped to manage the complications she suffered. She died 19 days later, leaving her husband, mother, and daughters to pursue a case that was repeatedly closed and reopened.¹⁰

Survivors built the evidence the program did not collect

People affected by the program organized locally and nationally, often across distance, language, illness, and limited income. AMPAEF formed to demand truth, justice, and comprehensive repair. The Quipu participatory archive enabled people in remote areas to record and hear oral testimony by telephone, organizing accounts around the program, the operations, life afterward, and the search for justice. A quipu here is not simply a metaphor for data; it reverses who gets to make the record and which consequences count.¹¹

Survivor advocacy had already produced a formal precedent in the case of María Mamérita Mestanza Chávez, a rural woman who died after a 1998 tubal ligation. In a 2003 friendly settlement before the Inter-American Commission, Peru recognized international responsibility, agreed to compensate and support her family, promised investigation, and accepted policy changes concerning consent, clinical conditions, reflection time, complaints, and staff training. The agreement was a state acknowledgment in one case and a program of promised reform, not a judicial finding about every operation or proof that the promises were implemented.¹²

Survivor organization also exposed the limits of a purely clinical explanation. Similar accounts across regions pointed toward campaign structure, target pressure, discriminatory ideas, and failed oversight. In 2024, the UN Committee on the Elimination of Discrimination against Women found that Peru violated the rights of five rural women who had been sterilized without informed consent. The Committee accepted their coherent accounts and the significance of their entry in Peru's registry while also recording Peru's position that the program was a general reproductive-health policy, not a discriminatory rule, and that registry entry was not an admission of state liability.⁸ The decision is powerful case-specific evidence; it does not adjudicate every procedure carried out under the program.

Peru created the Registry of Victims of Forced Sterilization in 2015 to identify affected people and connect them with legal, health, and psychological services. The Ministry of Women reported that its emergency centers had attended 6,103 registered cases and recorded 70,596 service activities from 2016 through September 2021. In the same official account, AMPAEF leaders Rute Zúñiga and Consuelo Tapullima described continuing ill health and delays in justice and reparation. Service counts show an institutional response; they do not measure care quality, coverage, recovery, compensation, or survivor satisfaction, and registry entry is not itself an adjudication of liability.¹³⁸

A court judgment makes consent an institutional obligation

In a judgment dated November 25, 2025 and notified on March 5, 2026, the Inter-American Court held Peru responsible for the forced sterilization and death of Celia Ramos Durand, the unjustified investigative delay, and harms to her family. It found violations of rights including life, integrity, private and family life, information, equality, and health.¹⁴ The Court ordered investigation; compensation and care; a unified consent instrument; updated protocols; patient guidance in official languages; complaint channels; a national sexual- and reproductive-health policy; inclusion in the survivor registry of people who died; and a reparations policy designed with victim participation.¹⁵

The judgment did not itself complete accountability. It recorded that the domestic investigation had still not produced a final determination, and it ordered Peru to proceed without using limitation rules or similar devices to shield grave rights violations.¹⁴ A Court status sheet updated in May 2026 listed 15 reparative measures as pending. That is the latest formal Inter-American compliance status confirmed here as of July 14, 2026; implementation after that date may change it.¹⁶

The harm also survives in people's relationship to public institutions. A 2025 peer-reviewed economics study compared localities with and without registered victims before and after public disclosure of the abuses. It estimated persistent reductions in contraception, prenatal and delivery care, and demand for public health services, along with worse child-health outcomes and lower trust in public institutions. The difference-in-differences design and robustness tests support a causal interpretation, but exposure is a district-level proxy built from a survivor registry rather than a case-level measure of what every resident knew or experienced.¹⁷

Target attainment was not neutral execution. Access measures remain necessary: a health system should know where services are absent, who waits, which methods are available, and whether outcomes are safe. But consent is not another metric to optimize after volume. It is a decision right held by the patient, including the right to leave the clinic without helping the institution meet its goal.

Brazil's Family Health Strategy places community health workers close to households too, but treats continuity, trust, and passage through care as the organizing problem. Peru's campaign shows what proximity becomes when a central target turns the worker's return visit into pressure for an irreversible procedure.

Connections and an organizational map

Four connections do different analytical work. Brazil's Family Health Strategy is a comparative contrast in primary-care proximity, not an asserted influence. The measurement, accounting, and control connection identifies the substitution of countable procedures for consent quality. Authority, legitimacy, and acceptance locates the conflict between state and professional authority and a patient's right to decide. Governance, stewardship, and accountability follows responsibility across executive direction, regional management, frontline action, investigation, courts, and survivor-governed repair. These are typed comparisons and interpretive lenses grounded in the cited histories, not claims that one institution caused or copied another.¹⁸

Six primary mechanisms

The program's declared reproductive-rights mission and its poverty-reduction logic place purpose, mission, and institutional legitimacy at the center. Unequal encounters between officials, professionals, and patients make authority, legitimacy, and acceptance equally central. Campaign pressure narrowed decision-making, judgment, and bounded rationality at the clinic door, while procedure counts made measurement, accounting, and control the operating language visible to senior officials. Clinicians' capacity either to protect or override refusal makes knowledge, expertise, and professional autonomy decisive. The long route from complaint to investigation and reparation makes governance, stewardship, and accountability the final primary mechanism.²¹⁵

Seven strong contextual mechanisms

The chain from presidency and ministry through regions, facilities, teams, and patients illuminates delegation, decentralization, and responsibility, coordination, communication, and common understanding, and structure, hierarchy, and scale. Targets and employment pressure affected cooperation, incentives, and organizational equilibrium, while recruitment, surgical campaigns, consent routines, and reporting made work design, productivity, and automation material even without digital automation. Survivors' descriptions of professional status, shame, solidarity, and organizing develop culture, informal organization, trust, and voice. Finally, completion data that omitted comprehension, pressure, complications, and lived consequences exemplify organizational ignorance: the system could know its throughput while failing to know whether the underlying decision was free.⁶⁷⁹

Mechanisms the public record develops only weakly

Later reviews, consent reforms, registries, and court orders provide limited evidence of learning, quality, and reliability, but adoption is not demonstrated implementation. Monthly reports and comparison of regional performance offer some evidence of executive attention, information, and organizational sensing, although the record does not reconstruct what each leader knew at each date. The assembled evidence does not materially develop strategy, competition, and adaptation or innovation, entrepreneurship, and renewal. Their zero scores mark evidentiary boundaries, not proof that adaptation or innovation was absent from all practice.¹⁸

The strongest organizational safeguard is therefore not a better consent form by itself. It is a system in which refusal counts as protected care; information is given in a language and form the person understands; irreversible procedures require time and independently reviewable consent; workers can report pressure without retaliation; adverse outcomes reach leaders as quickly as throughput; and survivors govern repair. Several of those controls now appear in the Court's orders; their effectiveness depends on implementation, monitoring, and survivor power rather than formal adoption alone.¹⁵ The moral test of a reproductive-health program is not how efficiently it changes a population statistic. It is whether each person remains the author of what happens to their body.

The assembled public record establishes central targets and incentives, documented consent failures, particular survivors' accounts, and authoritative findings in named cases. It does not establish one definitive number of forced sterilizations, every official's knowledge, a worker-level account of pressure and refusal, partner-by-partner financing and control, or completed accountability and repair. Publicly accessible survivor testimony is selective rather than an exhaustive representation of regions, languages, genders, or experiences. No reviewed evidence establishes material effects on nonhuman life or ecosystems. Those are research boundaries, not evidence that unmeasured consequences did not exist.