How Many Abortions Are Medically Necessary: The Data

No single, reliable number captures how many abortions are medically necessary, because no national surveillance system tracks that category and the medical community has never agreed on a universal definition of what “medically necessary” means in this context. The question sits at the intersection of clinical medicine and legal language, and those two worlds define the term differently. What the data does show is a wide range of conditions, from ectopic pregnancies to severe cardiac disease to aggressive cancers diagnosed in the first trimester, where ending a pregnancy is part of standard medical care to prevent the pregnant person from dying or suffering permanent organ damage. The gap between the political debate and the clinical reality is worth understanding in detail.

Why No Clean Number Exists

The most honest answer to “how many?” is that the data infrastructure to answer the question does not exist. In the United States, the CDC’s abortion surveillance system collects aggregate counts from state health departments, but it does not require states to record the medical indication, if any, behind each procedure. Hospital discharge data use diagnostic codes that were designed for billing, not for research into why a pregnancy ended. One validation study found that the ICD-9 code for missed abortion correctly identified actual cases of stable early pregnancy failure only about 42% of the time, meaning even the coding of miscarriage management is unreliable.

¹

The definitional problem runs deeper than record-keeping. “Medical necessity” in clinical practice is a judgment call made by a physician weighing a specific patient’s condition, comorbidities, and prognosis. A pregnancy that one clinician manages expectantly might be one another clinician would terminate, and both could be practicing within the standard of care. A paper examining the concept of medical necessity in the abortion debate noted that nearly half a century after Roe v. Wade placed the term at the center of public discourse, the US scientific community has still not produced a transparent, evidence-based framework for it.

²

What we can do is look at the conditions where the clinical evidence is clear that continuing a pregnancy poses a serious threat to the pregnant person’s life or health, and at the self-reported reasons women give for seeking abortions. Neither approach yields a single number, but together they sketch the landscape.

Ectopic Pregnancy and Incomplete Miscarriage

Ectopic pregnancies, where a fertilized egg implants outside the uterus (almost always in a fallopian tube), are the most unambiguous category. These pregnancies are never viable. Without treatment, they can rupture and cause fatal internal bleeding. Ectopic pregnancy is the leading cause of maternal death in the first trimester, accounting for roughly 5% to 10% of all pregnancy-related deaths.³ Treatment involves either medication (methotrexate) or surgery to remove the pregnancy, and no mainstream medical organization considers this optional. In the U.S., ectopic pregnancies occur in roughly 1 to 2 out of every 100 pregnancies.

Incomplete miscarriage is another situation where the same procedures used in elective abortion become medically necessary. When a pregnancy fails but tissue remains in the uterus, the standard treatment is to evacuate it, either with medication, manual vacuum aspiration, or dilation and curettage. Without that intervention, the patient faces serious risk of hemorrhage and infection.⁴ A study of 200 women with first-trimester miscarriage found that both manual vacuum aspiration and conventional dilation and curettage achieved complete evacuation in all cases, though vacuum aspiration had lower complication rates.⁵ These procedures are identical in technique to those used in elective abortion, which creates real confusion in legal environments where abortion is restricted. Physicians in restrictive states have reported uncertainty about whether managing a miscarriage could expose them to prosecution.

Preterm Membrane Rupture and Sepsis Risk

When the amniotic membranes rupture very early in pregnancy, before the fetus has any chance of surviving outside the womb, the pregnant person faces a prolonged waiting period during which infection can set in. This is called previable preterm prelabor rupture of membranes, or PPROM. Maternal sepsis occurs in up to about 7% of these cases, and rates are higher when patients are managed expectantly (waiting to see if the situation resolves) compared with when the pregnancy is ended.⁶ The Society for Maternal-Fetal Medicine recommends that all patients with previable PPROM be offered abortion care, and that expectant management should only be offered in the absence of contraindications.⁷

This is a condition where the distinction between “medically necessary” and “elective” breaks down entirely. The fetus has no realistic chance of survival. The pregnant person faces a meaningful risk of life-threatening infection. Waiting can be the more dangerous option, yet in jurisdictions with strict abortion laws, providers sometimes feel compelled to wait until the patient is actively deteriorating before intervening.

Severe Preeclampsia and Cardiac Conditions

Preeclampsia, a condition involving dangerously high blood pressure during pregnancy, normally develops later in gestation and is resolved by delivery. But when severe preeclampsia or eclampsia (seizures) develops before 24 weeks, delivery of a viable infant is not an option. A case series of 11 women who underwent abortion before 24 weeks due to severe preeclampsia or eclampsia found that none experienced severe complications from the procedure itself.⁸ Without ending the pregnancy, these patients face stroke, organ failure, and death.

Pulmonary hypertension, a condition of elevated blood pressure in the arteries of the lungs, carries some of the highest maternal mortality risks of any pregnancy complication. A meta-analysis found that severe pulmonary hypertension was associated with more than five times the maternal mortality risk compared with mild-to-moderate cases.⁹ Women with pulmonary hypertension experienced major adverse cardiac events at a rate of about 25%, compared with 0.4% among women without heart disease or pulmonary hypertension.¹⁰ Among women with congenital heart disease-associated pulmonary hypertension, those with moderate-to-severe disease had a maternal mortality rate of about 6%, compared with 0.1% for those with no pulmonary hypertension.¹¹ Many cardiologists counsel patients with severe pulmonary hypertension against pregnancy altogether, and when pregnancy occurs, termination is often recommended to prevent cardiac collapse.

Cancer Diagnosed During Pregnancy

Cancer diagnosed during the first trimester presents one of the more agonizing intersections of pregnancy and medical necessity. The drugs used in multi-agent chemotherapy regimens carry a high risk of congenital malformations, spontaneous abortion, and fetal death when administered in the first trimester, and clinical guidelines strongly recommend pregnancy termination in early pregnancy when aggressive treatment is needed.¹²¹³ The calculus shifts somewhat in the second and third trimesters, when certain chemotherapy regimens can be administered with lower fetal risk, but for cancers diagnosed early that require immediate, aggressive treatment, delaying therapy to preserve the pregnancy can be fatal for the patient.

This applies across tumor types. A case report of a pregnant woman diagnosed with primary diffuse large B-cell lymphoma of the breast described termination of the pregnancy followed by six cycles of intense multi-agent chemotherapy and radiation, resulting in complete response.¹⁴ Autoimmune diseases can also flare dangerously during pregnancy. Severe lupus nephritis flares in early pregnancy prompt discussions about therapeutic abortion as the safest way to manage the condition, though many women opt to continue the pregnancy and treat the flare directly.¹⁵

Fatal Fetal Anomalies

Some pregnancies involve fetuses with conditions so severe that survival outside the womb is impossible or will last only hours. These include anencephaly (absence of most of the brain), bilateral renal agenesis (absence of both kidneys), and certain chromosomal disorders like trisomy 13 and trisomy 18 in their most severe forms. An Irish study examining perinatal deaths found that about 36% involved a congenital anomaly, and among those anomalies with enough data to classify, roughly 42% were classified as fatal fetal anomalies.¹⁶

Many of these anomalies are not detectable in the first trimester. A study examining the timing of structural anomaly diagnosis found that only about 13% of fetal anomalies belonged to the category “always detectable” in the first trimester, while a quarter were classified as “never detectable” at that stage.¹⁷ This means that even with early screening, many fatal anomalies are only discovered at the anatomy scan around 18 to 20 weeks, or later. When families receive a diagnosis of a fatal anomaly after the first trimester, abortion may be sought not because the pregnancy threatens the mother’s physical health in the acute sense, but because carrying a pregnancy to term when the infant will not survive represents a different kind of medical and psychological burden. Whether this qualifies as “medically necessary” depends entirely on which definition you use, and that ambiguity is exactly why counting is so difficult.

Multifetal Pregnancy Reduction

Pregnancies with three or more fetuses carry dramatically elevated risks for both the pregnant person and the fetuses. Multifetal pregnancy reduction, a procedure that reduces the number of fetuses to improve outcomes for the remaining pregnancies, is another category where abortion becomes part of standard care. A Japanese prospective study examined cases where the indications were triplet or higher-order pregnancies, as well as twins with serious maternal complications.¹⁸

A population-based cohort study quantified the benefits. Compared with singletons, higher-order multifetal pregnancies without reduction had more than four times the risk of severe maternal morbidity and about ten times the risk of severe neonatal morbidity. Reducing a higher-order pregnancy to twins lowered those risks substantially, and reducing to a singleton brought the maternal morbidity risk close to baseline.¹⁹ These reductions are particularly common after fertility treatments, which carry a higher chance of multifetal implantation.

What Self-Reported Data Tells Us

One of the few data points that directly addresses the “how many” question comes from surveys of people who have had abortions. A study analyzing why women sought abortions in the United States found that about 12% mentioned health-related reasons. That figure broke down further: about 6% cited concern for their own health, about 5% cited concern for fetal health, and about 5% cited drug, tobacco, or alcohol use during the pregnancy (categories overlap because respondents could name more than one reason).²⁰

But this 12% figure understates the true scope of medically necessary abortion for several reasons. It reflects only what respondents chose to name as their primary or contributing reason, and a woman with both economic and health concerns might only mention the economic ones. It does not capture ectopic pregnancy treatments or miscarriage management, which are clinically identical to abortion but are often not classified or experienced as “getting an abortion.” And it does not capture cases where the life-threatening nature of the condition only became clear after the pregnancy was already further along. Suicide, for example, accounts for about a fifth of postpartum deaths, and perinatal mental health crises severe enough to be life-threatening are rarely counted in abortion statistics even when they factor into the decision.²¹

How Legal and Institutional Barriers Distort the Picture

The data on medically necessary abortion is shaped not just by biology but by the systems through which care is delivered or denied. In the United States, states that enacted abortion bans with no health exception saw a substantial increase in federal emergency-care violations. Specifically, these states experienced an additional 1.18 obstetric-related EMTALA violations per quarter, a statistically significant jump. Texas, which implemented its restrictive law in 2021, showed a similar increase.²² EMTALA requires hospitals to stabilize patients in medical emergencies regardless of other considerations, so a rise in violations suggests that patients with pregnancy emergencies are being turned away or experiencing delayed care.

Institutional restrictions predate the recent wave of state bans. Physicians working in Catholic-owned hospitals have described cases where abortion was medically indicated by their clinical judgment, but the hospital ethics committee delayed or denied the procedure as long as a fetal heartbeat was present, even when the patient was hemorrhaging and fully dilated at 17 weeks.²³ A separate study found that religiously affiliated hospitals varied widely in how they interpreted their own policies, and that institutional discomfort with abortion sometimes influenced decisions as much as formal religious teachings did. Patients encountered delays, financial obstacles, and restrictions on treatment as a result.²⁴

The problem is not limited to the United States. In Germany, where abortion after 14 weeks requires a physician to certify a medical indication of serious risk to the pregnant person’s life or health, the lack of national clinical guidelines has produced inconsistent interpretations of what counts. Providers described opaque criteria, stigma, and absent trauma-sensitive care resulting in delayed or denied treatment.²⁵ These barriers mean that some abortions that would be medically necessary under one definition simply never happen, or happen later than they should, inflating the risk to the patient.

The Broader Maternal Morbidity Context

To understand why the question of medically necessary abortion matters, it helps to understand the baseline level of danger in pregnancy. In the U.S., severe maternal morbidity during delivery-related hospitalizations has been increasing. Adjusted rates rose from about 147 per 10,000 discharges in early 2008 to about 180 per 10,000 discharges by the end of 2021.²⁶ Globally, a meta-analysis estimated the prevalence of severe maternal morbidity at about 2.5% of pregnancies, with maternal near-miss events occurring in about 1.7%.²⁷ A near-miss event is exactly what it sounds like: a situation where the patient almost died.

Not all of these morbidity events involve pregnancies where abortion was the indicated treatment, but the numbers make clear that pregnancy is not a benign condition. When a patient enters pregnancy with a serious preexisting disease, or when a life-threatening complication develops early, the clinical calculus of continuing versus ending the pregnancy is not theoretical. The researchers and institutions identifying legal restrictions, provider bias, and stigma as determinants of maternal mortality among people seeking abortion care are pointing at a measurable link between restricted access and worse outcomes.²⁸ The inability to produce a clean count of medically necessary abortions is itself a data failure with consequences for patients.