Pharmaceutical Freedom: Why Patients Have a Right to Self-Medicate

If Patients Can Refuse Care, Why Can't They Access It?

A Psychiatrist’s Read of Jessica Flanigan's Pharmaceutical Freedom

---

I. Opening

A patient once told me he trusted me completely with his mental health, but not with his prescriptions.

He said it casually, almost apologetically, the way someone might admit to a small embarrassment. "I trust your judgment," he clarified. "I just don't trust the system you're stuck in."

He had been taking alprazolam for years. Not prescribed by me. Not prescribed by anyone, technically. He obtained it from a pharmacy across the border during work tips without much friction, and he mentioned this the way someone might mention buying vitamins online.

What he wanted from me wasn’t access. He already had that. He wanted help stopping.

Not immediately, and not dramatically. He was not asking for a referral to a medical detox facility. He wanted a plan that respected the fact that the drug was already part of his life, a taper that preserved his job, his reputation, and whatever fragile equilibrium he had built around a substance he no longer fully trusted but still relied on.

In the standard version of psychiatry, this is a straightforward clinical problem. In the one we actually practice, it is something else. If I prescribe him alprazolam, I assume legal and professional liability for everything that follows. If I refuse, he continues taking the same drug anyway, outside the medical system, where I have less visibility and less influence. The policy regime under which we both operated was not preventing him from self-medicating. It had simply decided that the self-medication would happen outside our relationship rather than inside it. He had pharmaceutical freedom already.

Jessica Flanigan's book is about the question that follows. What does the world look like when my gatekeeping role as prescriber is removed? She asks it with more theoretical seriousness than most clinicians ever do, and her answer is more correct than I am comfortable with. The gap between her answer and the structure I work inside is where the interesting problem lives.

II. The Strange Idea

Pharmaceutical Freedom opens with two patients. Debbie has life-threatening diabetes and the legal right to refuse her insulin, even if refusal will kill her. Danny's diabetes is milder; he wants insulin because he wants nothing to do with lifestyle change, and he does not have the legal right to obtain it without permission. Flanigan's whole book unfolds from a single observation about that pair: the same competence that licenses Debbie's refusal also licenses Danny's request, and there is no clean philosophical reason to honor one and deny the other.

The asymmetry has a legal origin. In Schloendorff v. Society of New York Hospital (1914), deciding in favor of a woman who awoke from anesthesia to find her uterus removed, Justice Cardozo wrote that every adult of sound mind has a right to determine what is done with his own body, and that a surgeon who operates without consent commits an assault.^[1] American medical ethics has spent the century since taking that line seriously on the refusal side and quietly ignoring it on the access side.

Flanigan's thesis is the line drawn through the asymmetry:

My thesis is that the same considerations that justify rights of informed consent also justify rights of self-medication. Paternalism is just as wrong at the pharmacy as it is in the doctor's office.^[2]

She runs the argument on two tracks. The first is epistemic. The people with literal skin in the game are usually the best judges of which drugs and side-effect profiles fit their lives, because the question of whether a drug is worth taking is not the same as the question of whether it is pharmacologically active. Drug safety, on this account, is a normative judgment dressed in the language of expertise:

Drug safety is not a scientific judgment that requires medical training and expertise about the physiological effects of a chemical. Drug safety is a normative judgment that requires knowledge about how the risks and side effects of a drug fit into a patient's life as a whole.^[3]

Physicians are experts in the biology. Patients are experts in what trade-offs are worth making. Both kinds of knowledge are required for any real decision about whether a given drug, at a given dose, is worth its costs. The current system asks the physician to authorize a decision that depends on information only the patient holds.

The second track is rights-based. Even when patients are poor judges, they retain authority over their own bodies; the right does not vanish because it is exercised badly. Liberal societies tolerate enormous self-regarding risk in other domains, including the most consequential decision available, which is the decision to refuse care altogether. The exemption carved out for pharmaceuticals rests, per Flanigan, on no deeper principle than regulatory history.

Her proposed alternative is not deregulation in the libertarian-cartoon sense. Flanigan envisions a certificatory system: drugs evaluated, labeled, graded for confidence and risk, and made available with that information attached. Physicians would still advise. Agencies would still publish their assessments. What changes is the location of the final decision. If officials treated drugs more like other products and treated patients as capable trustees of their own health, people could still consult medical experts and agency recommendations, but the decision to access any medication would be their own.^[4]

One question the framework appears to leave open is who pays. In an insurance-mediated system, the worry is that gatekeeping migrates rather than disappears, with the payer replacing the prescriber. Flanigan does address it, through collective bargaining by private drug plans, public vouchers for low-income patients with life-threatening conditions, and insurers retaining the right to condition reimbursement on consultation or certified-drug use.^[5] Whether that solves the gatekeeping problem or merely relocates it is a question I will return to in Section V. A second implication, noticed by several reviewers, is more uncomfortable: if her logic dissolves prescription requirements, it puts the same pressure on medical licensure, since both rest on the state authorizing certain people to grant access.^[6]

This acknowledgement of the incentives of physicians and regulators tees up one of Flannigan’s most provocative lines in the book:

Public officials kill people by enforcing prescription requirements.^[7]

Read quickly, this sounds like a utilitarian body-count claim: the deaths from delayed access exceed the deaths from premature access, therefore the policy is wrong. Flanigan is making a different and more interesting move. She is invoking the philosophical distinction between killing and letting die, and arguing that when a state actively prevents a competent person from obtaining a drug they would otherwise be able to use, the state has crossed from omission into action. The death does not register as the regulator's bad luck. It registers as the regulator's deed. Whether one accepts the move depends on prior commitments about agency and causation, but the claim is deontological in form, not consequentialist, which is where most debates about the FDA’s “hidden graveyard” often sit.

When I was first exposed to Flanigan, during an interview with Julia Galef several years ago, the thesis felt radical. It took only a few years of psychiatric practice to realize that it was also descriptive.

III. The Shadow System

The first thing to say about pharmaceutical freedom is that motivated patients already have it.

I see this constantly in clinic, in patterns that have hardened into a kind of informal infrastructure no one designed and no one supervises. In addition to the benzodiazepine case in the introduction, a few archetypes appear most often.

The first is the patient who continues a medication originally prescribed somewhere else, often abroad, refilled through channels that operate adjacent to the formal system. One of mine is an engineer who was prescribed an antipsychotic called amisulpride years ago by a psychiatrist in another country and simply kept the treatment when he moved. The medication is filled overseas, mailed to him, and taken without interruption. To him, nothing about the arrangement is illicit; he is continuing something that works, and the fact that the drug is not FDA-approved is an administrative detail. From my side it is more ambiguous. I am managing the effects of a medication I cannot prescribe, cannot source, and only partially see; responsible enough to worry about it, not empowered enough to regulate it.

The second is the high-functioning professional who arrives with a printout about modafinil, peptides, and a stack of nootropics ordered from a research-chemicals site. He is not asking permission. He is asking for calibration, a second opinion the seller could not provide, on a drug he has already obtained and intends to keep using.

The scarcity bottleneck for these individuals was not access, but guidance.

The second thing to note is that even within the regulatory frame, psychiatrists routinely bend the rules to expand access in at least two ways: diagnostic expansion and off-label prescribing.

If a medical diagnosis is required to access a pharmaceutical, this gives manufacturers and clinicians a quiet incentive to expand diagnostic categories that gate access. Each of the five iterations of the Diagnostic and Statistical Manual of Mental Disorders (DSM) from 1953 to 2013 approximately doubled the number of diagnosable mental “disorders.” Adult ADHD is the obvious example, with diagnosis rates shifting in ways that conveniently track demand for stimulants. Beta-blockers for performance anxiety folded into “social anxiety disorder,” modafinil for “shift work sleep disorder,” and Vyvanse for “binge-eating disorder” are others. Many of these clinical conversations comprise a collusion between doctor and patient to authorize access to drugs that, at their core, are more about lifestyle enhancement than treating diseases.

Flanigan also uses the existing practice of “off-label” prescribing as a real-world model to argue that strict premarket efficacy testing and FDA gatekeeping are unnecessary. Psychiatry's use of off-label prescriptions massively outpaces the rest of medicine. It is what happens when access requires a disease label and the label is the gatekeeper's instrument; what looks like nosology is, in part, an access architecture.

Flanigan's argument reads, from the clinic, less as a radical departure than as a formal recognition of an arrangement medicine already accommodates by accident, with the worst possible distribution of accountability.

If the question is whether prescription requirements protect patients from drugs they should not have, the answer in any individual case is unknowable, and the regulatory frame congratulates itself on the harms it cannot see. If the question is whether prescription requirements make it easier or harder for patients to get honest, well-informed clinical input on the medications they are already taking, the answer is unambiguous and points the wrong way.

IV. Where Psychiatry Strains the Argument

The pharmaceutical freedom argument has three places where it deforms under clinical weight, and they are related.

The first is that capacity, in the clinical sense, is not stable. Most of medicine treats decisional capacity as durable enough to function as a threshold; you have it or you do not, and once you have it your decisions are yours. Psychiatry sees capacity as task-specific, time-varying, and frequently altered by the very condition the patient is being treated for. The hypomanic patient asking for stimulants experiences his request as the most rational thing he has done in months. The severely depressed patient refusing care offers a coherent rationale built on a future from which she has been emotionally evicted. The patient who started a benzodiazepine for anxiety has, over two years, crossed into withdrawal-avoidance without noticing the transition.

Carroll, in her 2022 critique of Flanigan, makes a structurally related point from a different direction. She argues that Flanigan ascribes more competence to people as patients than as voters, and that the asymmetry is unstable.^[8] Carroll runs it as a political-philosophy puzzle. From inside the clinic the same instability is visible at much closer range, because we watch capacity decay and recover within a single relationship. Flanigan's response to Carroll deserves to be represented, because it is not the easy move it could have been. She concedes that patients are sometimes overconfident and that physicians and lawmakers are also overconfident; her claim is not that patients are uniquely rational, but that the incentive structure at the pharmacy differs from the incentive structure at the polls. A patient internalizes the consequences of her own pharmaceutical choices. A voter, casting one ballot among millions, does not. Voters, on her account, are distinctively prone to misjudging drug policy because the harms of restriction are dispersed and statistical while the harms of access are concentrated and televised, and loss aversion does the rest. I find this partly persuasive and partly an artful reframing. The empirical question of whose biases are more decision-relevant is not settled by stipulating which biases get to count.

In her 2022 response to her critics, Flanigan is unusually direct about what her framework cannot yet do:

Questions about risk are even trickier for non-consequentialists like myself, and I'm afraid I still don't have any satisfactory answers about how to apply deontological moral principles to cases where there is some risk of a catastrophe or some risk of a serious rights-violation.^[9]

This does not refute her argument. It acknowledges that psychiatry routinely operates in the territory she has conceded her framework cannot yet describe.

The second strain is what I have come to think of as the recursion problem. Some drugs do not merely carry risk; they alter the conditions under which subsequent decisions about themselves get made. The patient consenting at week one is not always the same agent as the patient consenting at week ten, because the faculty being asked to consent has been modified by the substance under consideration. Flanigan treats voluntary risk assumption as morally sufficient, on the model of any other free choice. Psychiatry has spent decades watching that assumption decay, sometimes slowly enough that the decay registers only in retrospect, when the patient looks back and cannot say when the line was crossed.

The third strain is that the boundary between self-regarding and other-regarding choice is more porous in psychiatry than in most domains. Kianpour has shown that Flanigan's exception for antibiotics, applied consistently, extends further than she allows; varenicline carries documented neuropsychiatric effects with downstream consequences for others, and alcohol's externalities are catastrophic.^[10] Sedatives affect driving. Stimulant misuse affects sleep, mood, and parenting. The clean self-regarding case turns out to be a smaller class of decisions than the argument needs it to be.

Quong, in the most rigorous philosophical engagement with Flanigan, argues from a different angle. The cleanest justification for the Doctrine of Informed Consent is bodily integrity, and bodily integrity does not obviously extend to a right of access.^[11] Refusal protects the body from invasion; access requires positive cooperation from the world. Flanigan's response is that bodily rights are themselves grounded in a more general right against interference, and that prescription enforcement is itself a form of coercion that interferes with what one does with one's body. Both views can hold. The force of Quong's wedge is to remind us that the symmetry argument is doing more philosophical work than it sometimes appears to be.

None of these tensions refute the argument, but they do locate where it starts to bend under the weight of clinical specifics, and they converge on a single variable her account keeps treating as fixed: the stability of the person being asked to assume the risk. That variable, not the risk itself, is where the operational problem lives.

V. The Third Person in the Room

The clinical encounter is not a two-party relationship. There is a third presence in the room, composed of malpractice precedent, DEA scrutiny, prior-authorization architecture, board-of-medicine expectations, and institutional risk management. It does not speak. It shapes what can be offered, what can be discussed openly, and what stays outside the bounds of the conversation. The patient senses it. The prescriber feels it more acutely.

Flanigan directs her critique at the state, as if the state were a discrete actor whose policies could be evaluated on their merits and replaced with better ones. The state is not in the room. Individual prescribers are in the room, and the policies of the state arrive there refracted through professional liability, employer policy, payer requirements, and the residue of every malpractice case any of us has read about. What looks like medical paternalism is, in many cases, professional survival adapting to a system in which the visible costs of one kind of error are concentrated and the visible costs of the other are dispersed.

The cost of error is asymmetrically distributed. A patient who refuses care and deteriorates is generally respected as having exercised autonomy; the death is sad, but it is hers. A prescriber who provides risky access and watches harm follow is scrutinized, second-guessed, sometimes sued, occasionally disciplined by their state medical board. These outcomes are not philosophically symmetric, and prescribers have absorbed that asymmetry into their reflexes. Daniel Carpenter, in his analysis of the FDA in Reputation and Power, names the same pattern at the regulatory scale: visible harms are minimized, dispersed harms are tolerated, and the institution's incentive structure runs in one direction.^[12] The clinic operates on the same logic as the agency, scaled to a single examination room.

There is something genuinely appealing to me about Flanigan's proposal from inside this structure. If physicians are not the gatekeepers, they are not the bearers of the consequences. The consultant role, in which I inform rather than authorize, is the more honest description of what I am actually competent to do. It removes a distortion in the encounter that the current regulatory structure quietly imposes. I would, in some moods, sign on tomorrow.

The cost of removing it is that gatekeeping organizes responsibility in ways that are not always coercive. The patient who can be conditionally offered a risky medication, contingent on a safety plan and a follow-up visit, is in a different conversation than the patient who can simply buy it at retail. Whether that difference is therapeutic or merely bureaucratic depends on the case, the patient, and the medication; the honest answer is that it is both, in proportions that vary.

Quong's distinction returns here in a form Flanigan does not fully address. Removing physician gatekeeping does not eliminate the coercive structure around pharmaceutical decisions. It relocates it. Insurers, employers, licensing bodies, fitness-for-duty examiners, family courts, child-welfare agencies: these are the secondary gatekeepers waiting to absorb the function. A pilot whose airline learns he has been self-medicating does not become more autonomous because his prescriber is no longer the bottleneck. Flanigan's payment proposals, which include collective insurance bargaining, conditional reimbursement, and vouchers, are sensible on their own terms, but they describe a system in which insurers retain decisive control over which drugs are practically affordable, and that control is itself a form of gatekeeping.

Her reform of the prescription system would not abolish gatekeeping. It would migrate it from a clinical relationship, where there is at least some individual accountability and some possibility of negotiation, into administrative structures with less of either.

VI. What the Book Gets Uncomfortably Right

This is where Flanigan's abstraction becomes clinically useful. There is a version of her argument that sounds exaggerated until you follow it through the structure described above. Delay accumulates in individual patients. In aggregate it looks like a policy. In the clinic it looks like a pattern.

Four things in the book survive clinical scrutiny better than I would like them to.

The first is that delay is a body count, and the count is not zero. Flanigan's most provocative line, the one about prescription requirements killing people, lands harder here than in the section that introduced it, because by now the reader has the structure to see what she means. The sequence is familiar to anyone who has watched a patient cycle through the standard-of-care options for a treatment-resistant condition. An early-phase signal is not strong enough to meet the chosen endpoint. The endpoint was chosen to withstand regulatory scrutiny rather than to register clinical benefit. The trial extends. Approval waits. The patient cycles through partial responses, accumulates side effects, loses jobs and relationships, and eventually either stabilizes on something inadequate or does not stabilize at all. The drug arrives later, with narrower labeling and higher evidentiary confidence. Some patients benefit. Others have already moved on, in one direction or another. No one counts the ones who did not wait.

This is the asymmetry Carpenter describes at institutional scale. The cost of being wrong in one direction is concentrated, identifiable, traceable to a decision; the cost of being wrong in the other direction is dispersed across a population that never appears in the same frame as the decision that produced it.^[13] That asymmetry constrains what I can offer the person sitting in front of me. The menu of available treatments at any given visit is a function of what is approved, what is labeled, what is defensible, and what is reimbursable. I am choosing among the survivors of a filtration process that selected for evidentiary confidence at the cost of timeliness, and for legal defensibility at the cost of clinical range. Flanigan is right that delay carries a cost. Carpenter explains why that cost is tolerated. The clinic is where the two positions meet and refuse to resolve.

The second is regulatory inconsistency. Alcohol is sold at gas stations. Tobacco is regulated at point of sale rather than at access. The supplement industry sells substances with measurable pharmacological effects under almost no oversight, some of which overlap meaningfully with prescription pharmacology. Meanwhile, drugs with established mechanisms, known dosing, decades of safety data, and clear therapeutic niches remain tightly controlled. The boundary does not track risk. It tracks regulatory history, which is to say it tracks the order in which different industries developed, captured their respective agencies, and stabilized their privileges.

The third is that restriction redistributes harm rather than eliminating it, and the redistribution is stratified by resources. The patients who already operate in a world of pharmaceutical freedom are overwhelmingly wealthy, educated, well-connected, and white. They have the time, language, money, and confidence to use international pharmacies, online vendors, supplement markets, ketamine-clinic networks, and direct-to-consumer telehealth. They self-experiment with peptides, pay cash for novel neuromodulation protocols, and design off-label regimens with help from physicians willing to advise them and from AI tools that explain pharmacokinetics on demand. One patient with bipolar disorder, able to afford a nonstandard neuromodulation course out of pocket, designed a variant protocol after reading the literature; it seems to have helped him, and the clinical work was free to proceed because he could bypass the insurance pathway entirely. Another patient, after years of severe fibromyalgia, researched a newly approved neurosteroid obsessively and had two weeks of striking relief on samples left at an office; her right-to-try appeal failed and she could not afford the cash price. What separated those two patients was resources rather than biology. Patients without those resources remain inside the formal system, where access is slower, narrower, and subject to authorization workflows designed by people who do not have to use them. Psychedelic therapy is the cleanest contemporary example: psilocybin retreats in legal jurisdictions, ketamine clinics, and underground guides are available to people who can afford them, while the same compounds remain federally inaccessible to the patients most likely to benefit and least likely to find a way around the prohibition.

The fourth is that restriction shapes honesty. This is essentially Anomaly's learned-helplessness point, and it deserves the most attention.^[14] When access depends on prescriber approval, the patient has a strong incentive to present in whatever way maximizes the likelihood of getting what they want. Symptoms are emphasized or hidden. Histories are edited. Substance use gets reframed as something else, or omitted. This is not lying in the sense that anyone would prosecute. It is adaptation to a system where the prescriber's authority over access creates a corresponding pressure on the information that prescribers receive. Anomaly's framing draws on Mill: state restrictions, by promising to manage risks on the citizen's behalf, can permanently stunt the development of the very faculties that would have allowed the citizen to manage them. The clinical version is more local. The frame that promises to protect the patient from bad decisions also produces the patient who cannot tell his doctor what he is actually doing.

A system that restricts access does not eliminate risk. It redistributes it: toward patients who cannot find their way around the workarounds, into clinical encounters where honesty has been priced out, and onto the timelines of patients waiting for permission that may never arrive. The conversation in which the options run out happens more often than the policy debate suggests.

VII. Toward a Capacity-Based Hybrid

The book's great virtue is that it forces clinicians to defend the gatekeeping role rather than assume it. Most pharmaceutical regulation is implicitly risk-based. The higher the perceived risk of a drug, the tighter the controls on access. Flanigan's most useful contribution, after the symmetry argument, is to demonstrate that risk on its own is a poor foundation for coercion; many activities of comparable or greater risk go entirely unrestricted, and the threshold at which paternalism becomes legitimate is not derived from any consistent principle.

A capacity-based framework offers a different organizing principle. The threshold for restricting access is not the level of risk involved, but the integrity of the decision-making process about that risk.

This is not a foreign idea in psychiatric ethics. The same logic has been argued in the literature on involuntary intervention, particularly around the question of whether dangerousness alone is sufficient justification for civil commitment, or whether decisional capacity should carry more of the moral weight. Extending that logic from involuntary hospitalization to pharmaceutical access is a smaller move than it appears, because it preserves the structural insight that what matters is not what the patient might do but whether the patient is the one doing it.

The right analogy is not the prescriber as gatekeeper but the prescriber as fiduciary advisor. Financial advisors do not authorize their clients' trades; they cannot prevent a client from making a foolish investment. What they offer is a relationship across time, a track record of trust long enough to make persuasion possible, and the authority that comes from being someone the client has chosen to listen to. The advisor's job is to deepen the conditions under which the client can exercise autonomy well, not to override it. Having no coercive authority is what makes the advisor freer to be honest about what the client is doing wrong. That structural freedom is precisely what the prescriber's role currently lacks.

It is worth being concrete about what that version of the visit looks like, because the appeal of the fiduciary model is easy to assert and harder to picture. The question shifts from "will I authorize this?" to "what would make this decision safer, more reversible, and more honest?" The clinical work becomes dose literacy, interaction review, sleep protection, monitoring against agreed-upon criteria, exit planning, and repeated attempts to keep the patient's future self in the conversation. None of that requires the prescriber to hold the keys. All of it requires the prescriber to still be in the room, which is exactly what the current arrangement makes difficult once the patient stops needing the prescription pad.

Two patients of mine illustrate where this works and where it does not.

The first is a high-performing professional whose stimulant regimen had become inseparable from his work demands. He came to me on a high dose of immediate-release amphetamine, sharp and effective in the morning, dysregulated and barely sleeping by night. I tried, repeatedly, to negotiate a regimen that would protect his cardiovascular health and his sleep. When I eventually drew a firm boundary because I believed the risks had become medically inappropriate and professionally indefensible, he rejected my recommendations and chose to continue pursuing a higher-risk approach outside my care. No one checked his blood pressure. No one encouraged him to sleep more than five hours. He was making a strategic and informed choice about his life, albeit one I believed carried substantial medical risk. What I could still offer him was risk management, harm reduction, longitudinal observation, and ongoing efforts at persuasion across months of follow-up. But under the current regime, our relationship became inseparable from my willingness to authorize the prescription itself. Once I could no longer ethically do that, the therapeutic alliance collapsed with it. The fiduciary-advisor role, the role that might have allowed continued monitoring and incremental risk reduction even amid disagreement, was no longer available, and he disappeared back into the same shadow system Section III describes.”

The second is a newly adult patient, recently turned eighteen, hypomanic, with just enough insight and just enough family pressure to keep showing up for weekly visits while deteriorating in front of me. He wants stimulants. In his case, gatekeeping is one of the few things slowing a trajectory toward something catastrophic. His insight is genuine but unstable; the hypomania is exactly the kind of state that makes his current preferences a poor guide to what a more regulated version of himself would endorse. A capacity-based framework does not free me to authorize what he asks for. It tells me to delay, to introduce friction, to keep the relationship open across a window in which his decisional capacity has a reasonable chance of returning. What protects him is the structural reason to keep him in a relationship long enough for the recursion problem to resolve in his favor, rather than a refusal issued in the abstract.

These two are the central cases the hybrid has to handle. The first is where current regulation fails, because it removes the option of honest fiduciary engagement, and the patient ends up worse off because the clinician could not be useful in the only register he needed. The second is where pure pharmaceutical freedom fails, because the chooser is the variable. Any framework that cannot tell the difference between these is too coarse to operate in psychiatry.

The default of a capacity-based system would be broad. Competent adults have wide latitude over what they take and why. Physicians serve primarily as advisors. Information remains central to the encounter. The final decision rests with the patient, in something close to the certificatory model Flanigan proposes.

The friction layer is the modification. Certain classes of medication, those that reliably alter the conditions of subsequent decision-making, are accessed through systems that introduce friction proportional to their capacity to disrupt autonomy over time. Not prohibition; friction. Structured initiation, time-limited supplies, scheduled reassessment, transparent monitoring against agreed-upon criteria. The goal is not to eliminate the risk, which is the patient's to take. The goal is to ensure that the decision to keep accepting the risk remains the patient's own, rather than the decision of an altered version of the patient that the substance has produced.

There is one consequentialist concession the hybrid has to make and Flanigan, as a deontologist, can decline to. Means restriction in suicide prevention is one of the cleanest empirical cases for friction-based access. Restricting access to highly lethal means probably saves lives at the margin, and the patients who die because they could not reach a particular method are not the same patients who might die because they could. The hybrid has to acknowledge this and either absorb the cost or explain how the friction layer accommodates it. It does some of both, imperfectly. Pretending otherwise would collapse into the same regulatory triumphalism the rest of this essay has been pushing against.

My view is therefore closer to Flanigan than to the status quo, but not as close as her argument requires. Competent adults should have broader pharmaceutical latitude than they have now. Psychiatry's exception is not risk; it is impaired or unstable capacity, especially when the drug itself can worsen the instability.

What this changes is the location of moral weight. Patients become more explicitly responsible for their own decisions. Physicians become more clearly advisory, which is closer to what we are actually competent to do. Honesty becomes possible because the incentive to manipulate access weakens; if the prescriber is no longer the bottleneck, there is less reason to lie to him. Capacity assessment becomes a continuous process rather than a one-time threshold, which matches how psychiatric capacity actually behaves. What it does not change is the difficulty of measuring capacity in real time, with incomplete information, under pressure. Any framework built on capacity inherits the limits of that inference, and Flanigan's 2022 concession about deontological risk applies here too. The hybrid does not resolve the underlying tension; it makes the trade-offs visible and locates them where someone can be held accountable for them.

The structural move is that the hybrid is not "some autonomy and some paternalism in proportions to be negotiated." It is a relocation of where the moral weight sits in the encounter. Flanigan asks whether autonomy is sufficient to justify access. The hybrid approach asks whether capacity is intact enough to make autonomy meaningful. These are different questions, and they yield different policies even when they yield similar outcomes in the easy cases.

The old question asks how much risk we are willing to tolerate in the name of freedom. The better question asks how much instability of the chooser we are willing to tolerate in the name of respecting them.

VIII. The Deeper Question

The patient in the opening had walked where the prescribing system could not follow. He did not need access. He needed someone willing to meet him where he had already arrived, and to attach guidance to a regimen the formal system had been pretending did not exist. The current arrangement, in his case, did not fail to grant him freedom. It failed to recognize the freedom he had already exercised.

That is the version of Flanigan's argument that is hardest to answer from inside the clinic. The structures she critiques are real. The harms she identifies are not abstractions. The regulatory frame she wants to dismantle is producing, in the patients I see, exactly the kind of distorted clinical encounter she predicts.

And yet the framework she offers does not survive contact with the kinds of patients for whom autonomy is the variable rather than the constant. The recursion problem is real. Capacity is unstable. The substances under discussion sometimes alter the faculty that justified their use. These are not edge cases in psychiatry; they are the central case. Flanigan's framework converts that fact into an externality and treats it as the price of respecting autonomy. The clinical answer treats it as constitutive of the practice and adapts. Neither is wrong. Each is incomplete in the direction the other is correct.

What a more honest version of the debate would acknowledge is that the question of access has quietly migrated. The patients who most embody Flanigan's argument are not waiting for her policy reform; they have already organized their lives around its absence, and the medical system encounters them in a posture of selective ignorance. AI tools now explain pharmacokinetics, suggest taper protocols, and reason through cross-tapers using the same evidence I am consulting in a side window during a telehealth visit. The information asymmetry that justified the gatekeeping has been eroding for a decade, and the regulatory architecture has not noticed. Medicine may not need to decide whether patients should have pharmaceutical freedom. Many already do. What it needs to decide is what it means to remain useful, clinically and ethically, in a world where the question of access has migrated to a different set of actors than the ones the regulations still address.

The patient who trusted my judgment but not my system understood this before I did. He had already built his own arrangement. He came back only for the part he could not build alone, which was a clinician willing to help him leave it. The structure we defend in public is not the structure we actually run, and the longer the description and the practice diverge, the more the defense starts to look like a way of not noticing the difference.