Lecture 5 — Autonomy, Paternalism, and Reproduction
Part A: Paternalism and Its Limits
0.1 A Demand in the Delivery Room
A woman in the second hour of labor is in extreme pain. She has had a dose of IV opioid that leaves her drowsy and slurring. Between contractions she grabs the rail, says she is “done with all of this,” pulls at the fetal monitor, and demands to sign herself out and go home. Her partner is frightened. The nurse believes she would not say this an hour from now.
Write or discuss:
- Should the team take her demand at face value and start discharge paperwork? If not, why not?
- What exactly is it about this refusal that makes overriding it feel safe — when overriding a calm patient’s refusal would not?
- How long may the team “wait her out” before stalling becomes a way of simply ignoring her?
Almost everyone would override here without hesitation. This chapter is about when that instinct is right and when it is a trap.
0.2 After You’ve Discussed
- Almost no one thinks we must take that demand at face value. We override people “for their own good” all the time — small children, a friend’s car keys, an unconscious patient in the ED.
- So the interesting question is never whether such overriding is ever justified. It is where it stops — and what separates the override we’d all accept from the kind that becomes tyranny.
0.3 Today
By the end of Part A, you will be able to:
- Define paternalism and state Mill’s harm principle.
- Distinguish soft from hard paternalism by whether the overridden choice is compromised or competent.
- Say where a state’s incentive becomes pressure, and where pressure becomes coercion.
- Place policies aimed at pregnant women on a soft-to-hard ladder, and state what Ferguson v. City of Charleston held.
- Explain deterrence-of-care, and why punitive prenatal policy produces worse outcomes.
0.4 A Puzzle to Hold
The US is among the most punitive peer countries on drugs, and has the worst outcomes.
- It imprisons people at ~541 per 100,000 — many times the rate of other wealthy democracies — and still has the highest overdose death rate of any rich nation (~21 per 100,000 in 2025, down from a peak near 33).
- Its infant and maternal mortality are also worse than its peers’.
- Yet on tobacco — taxed and age-limited rather than banned — its results are among the best.
Stringency alone does not explain this. Singapore and Japan are harsher than the US on drugs and have far fewer overdose deaths. The framework below is built to address the puzzle. (Tikkanen et al. 2022)
1 The Millian Framework
1.1 Paternalism, Defined
Paternalism is interfering with a person’s liberty, without their consent, only because you judge it to be for their own good (Dworkin 2020).
- The “only because” is what makes it paternalism.
- A seatbelt law that protected only the wearer is paternalism.
- A drunk-driving law is not: it protects other people on the road.
- Mill treats those two cases completely differently.
Why the “only because” matters. Many real laws have mixed motives. A cigarette tax protects the smoker (paternalism) and raises revenue and reduces secondhand smoke (not paternalism). When we ask whether a policy is “justified paternalism,” we are asking about the part of its justification that appeals to the chooser’s own good.
1.2 Mill’s Harm Principle
“The only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant.… Over himself, over his own body and mind, the individual is sovereign.”
— John Stuart Mill, On Liberty (1859), ch. 1
- Read strictly, this bans paternalism: the state may stop you from harming others, never only yourself.
- Mill’s own exceptions: children, and the man about to cross a bridge he doesn’t know is broken. You may stop him until he knows.
1.3 Mill’s Own Exceptions
The harm principle is not the all-purpose shield it is sometimes treated as. In the same book, Mill carved out who it does not cover.
Mill himself excluded:
- Children — “those who are still in a state to require being taken care of by others.”
- People who cannot be reached by “free and equal discussion.”
- His famous bridge: if a man is about to cross a bridge you know is broken and you cannot warn him in time, you may stop him — to make sure he knows.
What this tells us:
The case for leaving people alone rests on three assumptions — that the person is competent, choosing freely, and well-informed. Where those hold, Mill says hands off. Where they fail, even Mill lets us step in. The whole soft/hard distinction grows out of this.
1.4 Soft vs. Hard — The Real Dividing Line
The line is not how forceful the intervention is. It is what kind of choice you are overriding.
| Soft paternalism | Hard paternalism | |
|---|---|---|
| Overrides a choice that is… | compromised — coerced, ignorant, incapacitated, or impaired | competent, informed, and voluntary |
| What you’re really doing | restoring or protecting the person’s own agency | substituting your judgment for theirs |
| Mill’s verdict | permitted — it isn’t a real override of a sovereign choice | forbidden — this is the line he draws |
| The bridge | stopping the man until he knows the bridge is out | stopping him after he knows, because you think living is better |
Whether a given choice is compromised enough to count as soft is the judgment the hard cases turn on. (Feinberg 1986)
1.5 Is It Soft? Three Worked Calls
Clearly soft:
- The sedated, pain-flooded laboring patient from the opening. Capacity is transiently compromised; you wait it out to give the real chooser back. You are restoring agency, not replacing it.
- Age limits on alcohol, tobacco, and tattoos. The claim is that a 16-year-old’s capacity to weigh long-run risk is not yet whole — Mill’s “children” exception, drawn at a rough line.
Arguably soft — and contested:
- Cigarette taxes. A standard hard-paternalism reading says: a competent adult chose to smoke, leave him be. But addiction degrades the voluntariness of each cigarette. The smoker, asked when sober, usually wishes he smoked less. A tax pushes against the addicted choice and toward the reflective one — which looks more like soft paternalism in disguise.
1.6 The Three-Question Test
Before overriding anyone “for their own good,” ask:
- Is the person competent — able to understand and weigh the facts right now?
- Is the choice voluntary — free from coercion, manipulation, and (in the hard sense) from addiction or false information?
- Does the harm fall on the chooser alone, or also on others?
If the honest answers are yes, yes, and chooser alone, Mill says: leave the person be. A “no” on question 1 or 2 opens the door to soft paternalism. A “no” on question 3 takes the case out of paternalism entirely. The genuinely hard cases are the ones where every answer is yes and we still want to step in.
1.7 Where Is the Line?
A ladder of ways a state can push birth rates:
- A: free contraception and parental leave
- B: a $10,000 tax credit per child
- C: a $100,000 tax credit per child
- D: priority housing for families with three children
- E: tax penalties on childless adults over 30
- F: a ban on contraception (Romania, 1966)
- G: forced sterilization (China; the US, 1907–1979)
- Where does incentive become pressure, and pressure become coercion?
- Run step C and step E through the three-question test. What changes between them?
1.8 A Hard Case We Already Accept
You cannot buy morphine, oxycodone, or fentanyl without a prescription, even as a competent, fully informed adult. By Mill’s strict line that is forbidden hard paternalism. Nearly every country keeps it anyway.
- Addiction changes the chooser. The first dose is a competent choice; dependence degrades every dose after.
- Information is lopsided. Few patients can judge dose, interactions, and overdose risk.
- The harm is fast and irreversible. Mill’s case for liberty leans on learning from mistakes, and an overdose cannot be learned from.
Each defense works by showing the choice is less free than it looks. Much defensible hard paternalism takes that form.
1.9 Make It Contemporary
Live versions of the same gate, 2024–25:
- Australia’s prescription-only nicotine rule for vaping (in force 2024): you can vape, but you must go through a doctor.
- The FDA’s proposal to cut nicotine in cigarettes to non-addictive levels — gating the product, not the person.
- Tightened opioid-prescribing limits after the 2010s overdose wave — which then pushed some patients to street supply.
What each is really doing:
Each dresses hard paternalism in a clinical pathway. “You may have it — through a doctor” feels gentler than “you may not have it.” But it is still the state deciding a competent adult may not simply choose for himself.
And the opioid example warns of the cost: tighten the legal gate too hard, and you can push people toward something worse.
2 Why Drug Policy Is Hard
2.1 Two Different Questions
So far we have sorted overrides by whether they are legitimate — the soft/hard, Mill question. Whether a drug policy actually works is a separate question, and the answer is not “be softer.”
- Legitimacy (Axis 1): is overriding this choice justified? Turns on whether the choice is compromised or competent.
- What works (Axis 2): does the regime cut harm, and on whom does the burden fall? Turns on coherence, consistent enforcement, funded treatment, and supply — not on how harsh it is.
The puzzle from the start of class lives on Axis 2. Keep the two apart: a policy can be ethically defensible and still fail, or effective and still unjust.
2.2 The Case for Control: Unrestricted Opiates
Before modern drug laws, opioids were sold without restriction. The consequences were significant.
In 19th-century America, opium, morphine, and (after 1898) heroin were sold legally and without prescription. They were in patent medicines, “women’s tonics,” and over-the-counter remedies. Heroin was marketed by Bayer as a cough suppressant.
Infants were dosed with opium-laced “soothing syrups” — Mrs. Winslow’s among them — to quiet teething. Some died. By 1900, a large share of the addicted population were ordinary patients, mostly women, hooked through medicine they had been told was safe.
This is the case for paternalism in drug policy: a wholly unrestricted market produced mass addiction in people who never made an informed choice. (Young 1961)
2.3 The Case against Bans: Prohibition
The Eighteenth Amendment (1919) banned making, selling, and shipping alcohol; the Volstead Act gave it teeth. It was repealed in 1933 — the only constitutional amendment ever fully overturned.
In between, drinking fell briefly and then climbed back. Illegal liquor became a major industry, federal prison populations roughly tripled, and the murder rate hit a peak it would not match for sixty years.
Three lessons that recur in modern drug policy:
- Banning a widely-used substance does not erase demand for it.
- The black market that replaces a legal one is more violent.
- Enforcement falls hardest on the communities least able to push back.
Both poles — unrestricted access and total prohibition — produced serious harms. The policy question concerns what sits between them.
2.4 Stringency Doesn’t Predict Outcomes
| Country | Stance | Overdose (per 100k) | Incarceration (per 100k) |
|---|---|---|---|
| United States | Punitive | ~21 | ~541 |
| Portugal | Lenient | ~1 | ~110 |
| Netherlands | Lenient | ~1.5 | ~65 |
| Singapore | Very harsh | low | ~156 |
| Japan | Harsh | <0.5 | ~36 |
The lenient countries and the harsh ones both have far lower overdose mortality than the US. Stringency alone does not predict the outcome: the most punitive country in the table and two of the most lenient sit at opposite ends of the policy spectrum, and the US is worse than all of them. (Tikkanen et al. 2022; Institute for Crime & Justice Policy Research 2023)
A trend worth watching — and worth arguing about. US overdose deaths have now fallen for three consecutive years: roughly 110,000 in 2023, 81,000 in 2024, and 70,000 in 2025 — a drop of well over a third from the peak, and the lowest total since 2019 (National Center for Health Statistics 2026).
Nobody is certain why. Candidate explanations include the wide distribution of naloxone (now available over the counter), expanded access to buprenorphine and methadone, changes in the illicit fentanyl supply, and the grim possibility that the population most at risk has already been depleted. Each explanation implies a different policy lesson, and they are not mutually exclusive.
Write or discuss: Suppose the decline is mostly due to naloxone — a harm-reduction measure that makes drug use safer rather than rarer. Does that vindicate the harm-reduction model over the punitive one? Now suppose it is mostly supply-side changes in the fentanyl market. Does the same data now support a different conclusion? What does this tell you about reading a trend line as evidence for a policy you already favor?
2.5 Two Coherent Models
The countries that do well share not a stance but a coherent design — enforcement, treatment, and supply policy that pull the same direction. Two opposite designs both work.
Consistent, predictable strictness (Singapore, Japan).
Both states made deliberate, sustained efforts to keep drug use rare and supply thin: uniform enforcement, strong norms, and — in Singapore — compulsory treatment alongside heavy penalties. Use rates and overdose deaths are low, and have stayed low for decades.
Thoughtful decriminalization (Portugal).
Portugal removed criminal penalties for personal possession in 2001 and redirected resources into treatment, outreach, and dissuasion commissions. Overdose deaths, HIV transmission, and drug-related incarceration all fell substantially.
The US matches neither: punitive without consistency, and without the funded treatment that both models rely on. (Oregon’s 2020 decriminalization, repealed by 2024, removed penalties without building the treatment side — the Portuguese half it left out.)
2.6 What the Soft-Paternalism Middle Achieved
Between unrestricted sale and outright bans sit tools that moved the numbers without charging the user.
- Tobacco taxes, age limits, and ad and indoor-smoking limits. US adult smoking fell from ~42% in 1965 to ~11% in 2022; price accounts for about half the decline.
- The drinking age. Raised to 21 in 1984; alcohol-related traffic deaths among 18–20-year-olds fell by about a third over the next decade.
- Naloxone (Narcan). A widely distributed antidote, not a ban — the single step with the clearest effect on the modern opioid death rate.
The common mechanism is changing the price, availability, age of access, or downstream harm of a behavior rather than prosecuting those who engage in it.
2.7 Thought Question — Should the State Ban Tobacco?
Tobacco kills about 480,000 Americans a year — more than all illegal drugs combined. It is sold legally to anyone over twenty-one. The FDA has the power to cut nicotine to non-addictive levels; it has proposed this but not done it.
Write or discuss.
- Apply the three-question test to a thirty-year-old smoker. Is his daily choice really “competent and voluntary,” given addiction?
- Would you support a steep federal tobacco tax but not a ban? If so, what is doing the work in your view — and would the same reasoning apply to alcohol? To methamphetamine?
- If cutting nicotine to non-addictive levels is soft paternalism (it restores voluntary choice), is that a stronger justification than a ban — or a sneakier one?
2.8 After You’ve Discussed
A distinction worth carrying into the pregnancy block: deterrence-of-care.
An enforcement policy can drive the very people it claims to protect away from the system meant to help them. When tobacco rules target the user, smokers don’t quit — they hide it from their doctor. When drug enforcement reaches into pregnancy, women don’t stop using — they skip prenatal care.
The Axis-2 lesson lands on the clinician. The distinctive feature of US prenatal drug policy is less its severity than its structure: it routes enforcement through the clinical encounter — test, report, refer to police — placing the clinician within the enforcement chain. A state can target supply without doing so. The clinician’s question is therefore not “strict or lenient” but whether the clinical encounter should serve an enforcement function at all.
3 Pregnancy as a Paternalism Problem
3.1 When the Second Party Enters
- Mill’s third question, does the harm fall on the chooser alone?, assumes a single body. Pregnancy is the standard case where it does not.
- That creates pressure to reclassify a pregnant woman’s self-regarding choices as other-regarding, moving them outside the zone Mill protects.
- Poverty, addiction, an abusive partner, or no childcare often narrow her real options. That does not license overriding her, but it belongs in the judgment.
- The paired case, Adriana Smith, is the limiting case: declared brain-dead at nine weeks pregnant in Georgia in 2025, her body was kept on machines for four months over her family’s objection.
3.2 A Spectrum of Interventions
The aim, protecting the future child, is pursued with tools from mild to coercive.
| Intervention | Example | Type |
|---|---|---|
| Risk messaging | “No safe amount” of alcohol; broad deli-meat, sushi, soft-cheese cautions, sometimes overstating thin evidence | Soft (information-shaping) |
| Transient medical override | Waiting out the sedated, pain-flooded laboring patient’s demand to leave | Soft (capacity restored) |
| Mandatory reporting | Newborn toxicology → child-protective referral and a Plan of Safe Care | Mixed — services or surveillance, by state |
| Criminal prosecution | Prenatal drug use charged as child abuse or homicide | Hard — a charge pregnancy alone creates |
Climbing the ladder, the overridden choice gets more competent, so Mill’s case against interfering gets stronger, and the empirical record gets worse.
3.3 The Question, Stated Sharply
May the state charge a pregnant woman, on the basis of her pregnancy, with a graver crime than the same act would carry in anyone else?
- The conduct in the drug cases is already an offense. Outside pregnancy, methamphetamine use is at most simple possession; here it is charged as child abuse, chemical endangerment, or homicide.
- A fetal-harm theory is doing the work, and it is not limited to illegal drugs. It already reaches alcohol, prescribed medication, and refusing recommended care.
- The question takes no position on the moral status of the fetus. It asks about an enforcement practice and whom it reaches.
3.4 Ferguson v. City of Charleston (2001)
In the late 1980s, the Medical University of South Carolina in Charleston began a secret urine drug-testing program on some pregnant patients. Positive results went to police. Hospital staff wrote the rules; the local prosecutor signed off. Triggers included things like “no prenatal care.”
Ten of the eleven women arrested were Black. The eleventh was a white woman whose chart noted she “lives with her boyfriend, who is a Negro.”
Held, 6–3 (Stevens, J.): When the program’s main goal was to gather evidence for police, it was a Fourth Amendment search needing consent. The secret testing was unconstitutional.
Not decided: whether testing tied to treatment (no police) would be allowed, or whether civil steps short of arrest raise the same problem.
3.5 Tennessee’s Fetal-Assault Law — and the Data
2014. Tennessee made it a misdemeanor to take a narcotic during pregnancy “if her child is born addicted to or harmed by” it.
2014–16. About one hundred women were arrested, mostly in poor rural counties. Drug-use rates in pregnancy did not fall. Women reported avoiding prenatal care; some gave birth at home, alone, to dodge the hospital drug test.
2016. The legislature let it sunset.
What the national data show. Faherty et al. (2019), studying 4.5 million births, found states with punitive policies (criminal charges, civil commitment, or reporting as abuse) had a 27% higher rate of newborn drug withdrawal — the very outcome the laws were sold as preventing.
This is deterrence-of-care: a policy that drives the people it claims to protect away from the care meant to help them.
Neonatal abstinence syndrome (NAS) — newborn opioid withdrawal — is the set of symptoms (tremors, feeding trouble, sometimes seizures) a baby shows after in-utero opioid exposure. It is treatable, and most babies do well with good care. It is also the diagnosis punitive reporting laws claim to reduce — and don’t.
3.6 The Scale, and a Name
- Paltrow and Flavin counted 1,396 arrests and forced interventions on pregnant women in the US from 1973 to 2005: 71% low-income, 59% women of color. The most common charge involved drugs; the second, refusing care.
- In 2021 an Oklahoma jury convicted Brittney Poolaw, 21, of first-degree manslaughter for a pregnancy loss at about seventeen weeks. The medical examiner found no cause, and the state’s own toxicologist said the trace methamphetamine was not clearly the cause. About one in five pregnancies is lost on its own at that stage. She was sentenced to four years.
Read the number carefully. 1,396 is not the count of pregnant women who used drugs. It is the count who were charged, jailed, or forced into a procedure.
Dorothy Roberts’s Killing the Black Body (1997) explains why these patterns track race and class.
3.7 Where the Profession Draws the Line
ACOG Committee Opinions 473 / 664:
- Substance use in pregnancy is “a medical condition that should be managed with the patient.”
- Criminal charges and forced detention are “not effective in reducing” drug or alcohol abuse and are “contrary to the welfare of the mother and the fetus.”
- Testing without consent is rejected; CPS reporting “depends on state law.”
Why this is more than an opinion.
ACOG’s ethical claim (manage it as medicine, don’t punish) and the population data (Faherty: punishment → worse withdrawal rates) point the same way.
That kind of agreement between an ethical argument and the empirical record is rare enough to carry real weight.
(American College of Obstetricians and Gynecologists 2011, 2016; Faherty et al. 2019)
3.8 Thought Question — A Patient on Methamphetamine
A 32-week-pregnant patient arrives in early labor. Her urine is positive for methamphetamine. She has had three prenatal visits, no other problems, and says she used “yesterday, maybe the day before.” State law requires reporting drug exposure to CPS; the hospital may choose whether to involve police.
Write or discuss.
- Which rows of the pressure-spectrum are in play here, and which would you actually use?
- The attending asks whether to call CPS now or after delivery. What is the right answer, and why?
- Does your answer change if the drug is alcohol? Fentanyl? Cannabis that is legal in this state? Be honest about why.
- Since Dobbs, prosecutions for pregnancy outcomes have risen from about fifty a year to at least 412 in two years, nearly all for substance use (Pregnancy Justice 2025). Does that change what you do at this bedside?
3.9 Make It Contemporary
Three live trends, 2022–2025:
- Pregnancy Justice documented a sharp jump in prosecutions for pregnancy outcomes since Dobbs — from roughly fifty a year nationally to at least 412 cases in the two years following the decision (June 2022–June 2024), the highest totals it has recorded. The overwhelming majority — 399 of 412 — involved alleged substance use during pregnancy; 31 targeted people who had experienced a pregnancy loss, and 9 involved abortion allegations (Pregnancy Justice 2025).
- Since 2022, at least nine states have introduced fetal-personhood bills that would widen the basis for charges like Poolaw’s.
- Federal CAPTA still requires a Plan of Safe Care for every substance-exposed newborn — a service model that some states run as support and others as surveillance.
What unifies the punitive trend:
Each move treats the fetus as a legal subject whose interests limit what the pregnant person may do — pushing the “harm to others” exception to swallow Mill’s rule.
The medical position has not changed. The legal world it operates in has shifted sharply.
4 Partial Autonomy — The Hardest Case
4.1 When Autonomy Is Real but Incomplete
The framework has a clean answer at the two ends: full autonomy (hands off) and none (decide for them). The genuinely hard cases all live in the middle — where autonomy is partial.
We have already met three of them without naming the pattern:
- The addicted chooser — competent overall, but each drug choice is partly compromised.
- The pregnant patient — one chooser, but her choices reach a second party.
- And the clearest of all, the one Mill named first: the adolescent.
4.2 The Adolescent
A 16-year-old is not a small child and not quite an adult. The mature-minor doctrine lets some adolescents consent to some care on their own — and many states let minors consent without a parent to a defined list: contraception, STI treatment, mental-health and substance-use care. (Coleman and Rosoff 2013)
The logic is soft-paternalist on both sides: we restrict them (age limits, parental consent for surgery) because capacity is still forming, and we free them (confidential STI care) because forcing parental involvement would drive them away from care entirely.
Mature-minor doctrine — a legal rule, varying by state, that lets a minor judged mature enough to understand a treatment consent to it without a parent. It is decided case-by-case, usually for lower-risk care, and is the law’s attempt to draw Mill’s “children” line at something finer than a birthday.
4.3 Thought Question — A Fifteen-Year-Old’s Refusal
A 15-year-old with a serious but treatable cancer has a strong chance of cure with chemotherapy and almost none without it. He understands all of this clearly and calmly, and refuses — he has watched the treatment’s effects and does not want them. His parents are split. He is not, by any clinical measure, incompetent.
Write or discuss.
- Is his refusal more like the competent adult’s (honor it) or the small child’s (override it)? What in his situation pushes each way?
- If you would override him but not a 25-year-old in the identical clinical spot, exactly what is the difference doing the work — and is it the same thing that justified age limits earlier?
- Where would respecting his refusal be honoring his autonomy, and where would it be abandoning him?
4.4 The Cases, Sorted
| Case | Choice being overridden | Soft or hard? | Verdict |
|---|---|---|---|
| Sedated laboring patient | A demand made under transient incapacity | Soft | Override briefly, to restore her real choice |
| Prescription-only opioids | A competent adult’s purchase | Hard | Defensible, by showing the choice is less free than it looks |
| Reporting newborn drug exposure | A mother’s privacy | Mixed | Depends on whether the state runs it as services or surveillance |
| Criminalizing prenatal drug use | A pregnant woman’s conduct | Hard | Backfires: deterrence-of-care, worse outcomes |
4.5 The One Idea
Mill’s line is simple; life is not.
Leave competent, informed, freely-choosing adults alone. Step in only to protect others, or to protect a choice that is compromised — and then only as much as it takes to give the real choice back.
Every hard case is a fight over one word: voluntary.
Most defensible paternalism wins by showing the choice was less free than it looked — addiction, youth, transient incapacity, pressure. Most indefensible paternalism loses by ignoring what its enforcement does to the very people it claims to help.
4.6 Review Questions
Write a brief answer to each, or use them as discussion prompts.
Recall
In your own words: what separates soft from hard paternalism, and why does Mill permit one but not the other? Give one example of each.
Apply
A state proposes a steep tax on high-sugar sodas to fight diabetes. Use the three-question test to argue whether this is soft or hard paternalism — and whether that makes it easier or harder to justify.
Debate
Pick one: prescription-only opioids, the 21 drinking age, or criminal charges for prenatal drug use. Argue that the policy gets the paternalism question right or wrong, and say what should replace it.
4.7 Key Terms
- Paternalism
- Harm principle
- Soft paternalism
- Hard paternalism
- Three-question test
- Deterrence-of-care
- Neonatal abstinence syndrome
- Plan of Safe Care (CAPTA)
4.8 Further Reading
- John Stuart Mill, On Liberty (1859), ch. 1 — read it; it is short.
- Gerald Dworkin, “Paternalism,” Stanford Encyclopedia of Philosophy (2020).
- Joel Feinberg, Harm to Self (1986), on soft vs. hard paternalism.
- Lynn Paltrow and Jeanne Flavin, “Arrests of and Forced Interventions on Pregnant Women, 1973–2005,” JHPPL (2013).
- ACOG Committee Opinion 664 — read it.