New York's medical aid in dying law took effect on August 5, and it is the most demanding process of its kind in the country. A terminally ill adult who wants the medication must make an oral request that is recorded on audio or video and stored permanently in the medical record. They must submit a signed written form witnessed by two adults who are not relatives, heirs, or employees of the facility. Two physicians must independently confirm the six-month prognosis and the patient's capacity. A psychologist, psychiatrist, or neurologist must then evaluate the patient, a requirement no other state has. A five-day waiting period follows the prescription before it can be filled, and the patient must swallow the medication themselves, in a single conscious act. Each of these steps was added to protect someone from something, and each one is also a hurdle for the person the law was written for.
A philosophy professor recently argued in these pages that the mere availability of aid in dying burdens the dying with a new choice, turning continued life into something that requires justification. That is the philosophical version of the problem. New York's experience over the coming months will test the operational version: the choice exists, and now it must be scheduled, witnessed, evaluated, recorded, and timed, by a workforce that was already stretched before the law added its demands. The question is not only whether the choice burdens people. It is who can complete the process at all.
The process is the policy
The New York law did not invent the standard safeguards. Oregon's quarter-century-old statute established the template: a terminal diagnosis, two physicians, a written request, a waiting period. New York layered its own on top. The recorded oral request is unique among American states, a response to the fear that a family member or provider could claim a patient asked for the medication when they never did. The mandatory mental health evaluation is also a first, designed to catch depression or delusion that could masquerade as a settled wish to die. The witness restrictions are a treatise on coercion by themselves: the two signers cannot stand to inherit, cannot work for the facility treating the patient, cannot be the patient's health care agent. A person who pressures a patient into the process faces criminal liability.
Every one of those provisions answers a real objection that opponents raised. Cumulatively, they create a timeline. A dying patient must schedule at least three clinical evaluations with three different professionals, plus the recording, the witnesses, the form, the five-day wait, and the final act of self-administration. For a person with six months to live, six months of appointments is not a formality. It is the remaining life, spent in the very process the law created to keep them safe.
The state also decided, quietly, how the process ends on paper. The death certificate records the terminal illness, not the medication, and the law instructs that aid in dying is not to be counted as suicide. Those choices protect families from the stigma and insurers from the wording. They also mean the law's own use will be visible only through the provider reports the health department requires within five days of each prescription, which is how the state will learn what its process produced. The reporting system, like the rest of the law, is new, and the first year of data will be the first test of whether the machinery runs as designed or whether it selects quietly, the way administrative systems do.
The law assumes a workforce that is thin
The process runs on specialists, and the specialists are the shortage. Reporting in January found that the physicians who would guide many of these patients, neurologists in particular, are trained to manage disease rather than dying. Palliative care education in neurology residencies varies widely, and the specialty's pipeline for end-of-life care is thin. The gap is not abstract. Neurological patients are a disproportionate share of aid-in-dying requests nationally, and some of them face a biological deadline that outruns the paperwork. A person with ALS may lose the ability to swallow, which the law requires for self-administration, months before they lose the wish to choose. A California study found physicians wrote prescriptions for ALS patients a median of 76 days after the first inquiry, and in New York the calendar is crowded with additional steps that California does not require.
The mental health evaluation adds a separate bottleneck. Advocates flagged before the law took effect that psychiatrists and psychologists are scarce across upstate counties, and that arranging the evaluation on top of the attending and consulting visits could prove a serious barrier for patients with little time and less mobility. The evaluation exists to protect patients who are depressed or deluded. It also assumes that every county in the state has a mental health professional available, willing, and licensed to do this particular assessment on a dying patient's schedule, an assumption the state's own workforce data makes optimistic.
The friction selects, quietly
What a process like this does in practice is sort. It favors patients with family members who can drive to appointments, with health literacy high enough to navigate forms and recordings, with insurance that covers the visits, with physicians who participate at all, since participation is voluntary and Catholic-affiliated facilities and home hospice agencies may decline. None of that is written into the law. It is the law's operating system. The formal eligibility rules are simple: adult, resident, terminal, six months, capacity. The actual eligibility rule is whether the patient can complete a multi-visit administrative process in the time they have left, and the process takes longer precisely where the health system is thinnest.
Money adds one more unwritten rule. The visits, the evaluations, and the medication itself generally fall outside Medicare and most insurance coverage, a cost that lands on families already paying for end-of-life care. It is a small expense against the scale of dying, and it is also the only part of the process that can be resolved by a credit card. A law designed around a right available to every terminally ill resident ends up, in its first months, available in practice to those whose time, transportation, provider networks, and cash all hold out long enough. The state built the strictest version of this law in the country, and the strictness has a clientele.
The timeline has its own cruelty. The law was signed in February, the regulations arrived this summer, and some advocates worried out loud that patients who fought for the law would die before its machinery was ready. The five-day waiting period assumes five days to spare. The recorded request assumes a voice that can be recorded. The self-administration requirement assumes a body that can swallow. The law protects against haste from people who have months, and it asks haste from people who may not.
The two readings of friction
Every element of this design has two coherent readings, and the state's choice was to blend them. To supporters, the safeguards are the reason the law can be trusted: a strict process, freely entered, documented, and reversible, offers autonomy without abandoning the vulnerable. They point to Oregon, where the strict template has operated for a quarter century without documented cases of abuse. To opponents, the safeguards are a Trojan horse for normalization, and the proof is north of the border: in Quebec, where the process is wider, aid in dying now accounts for 7.9 percent of all deaths, the highest rate in the world, with three quarters of recipients over seventy. New York's Catholic health systems pushed back as the law took effect, declining to participate. This analysis takes no position on the contested questions about aid in dying itself. The observation that matters for the next year is narrower: the state's answer to the fear of too much death was a process that demands time, and the demand falls entirely on the dying.
The philosopher worried that offering the choice burdens the chooser. New York has built a system in which the choice, once made, becomes a schedule. Some patients will complete the schedule and die as they wished. Others will run out of time, or voice, or the ability to swallow, or the patience of exhausted families, and the law will have protected them from everything except being unable to use it. That is not an argument against safeguards. It is a description of what safeguards are: the process is the policy, and the policy meets the patient at the last appointment of their life. Whether the state has built a protection that works or a right that is too hard to reach will show up in the data the law itself requires, one prescription report at a time, and the first reports are only now being written.
Primary sources
- The New York State Department of Health announcement and the department's FAQ for the law's effective date, the recorded request requirement, the two-physician and mental health provisions, the waiting period, the witness restrictions, and the self-administration rule.
- The January STAT report on neuropalliative care for the clinician shortage analysis, the neurology training gap, the ALS timeline concern, and the 76-day California figure, and Courthouse News for the Catholic provider pushback.
- CBC's reporting for the Quebec rate and demographics, and the First Opinion essay by Monika Piotrowska for the burden-of-choice framing this analysis examines from the operational side.