American Health & Medical Research Association
Healthcare & informed decisions

Your right to refuse treatment, and its real limits

Adults who can decide are allowed to say no. We walk through how a refusal gets handled well, where the right stops, and why declining one thing isn't quitting care.

Sources14, listed below Reading time13 minutes
On this page
  1. What the right actually is
  2. Informed refusal is informed consent in reverse
  3. Refusing one thing is not refusing care
  4. Where the right stops
  5. The risk this site is not going to soften
  6. Walking out of a hospital early
  7. When a clinician takes it badly
  8. The single most useful thing here

Your doctor recommends a daily medication. You’ve read about the side effects, you don’t want it, and you say so. The room gets cooler. Someone uses the word “noncompliant.” You leave feeling like you did something wrong.

You didn’t. An adult who can understand their situation is allowed to turn down medical treatment, including treatment that would save their life. That’s not a loophole or a technicality. It’s the thing the entire consent system is built on, because a yes only means something if a no was available.

What most people don’t know is how much room sits between “yes” and “no.” You can decline one specific drug and accept everything else. You can say “not this year” and put a date on the calendar to revisit. You can ask what would have to change for the recommendation to change, which is often the most productive sentence in the whole conversation.

This article covers what the right to refuse is, where it genuinely stops, how a refusal gets handled well on both sides, and the honest risk of refusing effective treatment for something serious.

What the right is

MedlinePlus, the consumer health service run by the National Library of Medicine, states it directly: you may refuse treatment if you can understand your condition, your options, and the risks and benefits of each. A provider may recommend against your choice. A provider may not force treatment on you.

The American Medical Association’s Code of Medical Ethics says the same thing in its list of patient rights, Opinion 1.1.3: “A patient who has decision-making capacity may accept or refuse any recommended medical intervention.” Note the word “any.” The opinion also lists your right to hear about the option of forgoing treatment in the first place, and your right to have your decision respected once you make it.

Federal rules reach the same place from a different direction. The condition of participation on patient rights that hospitals must meet to bill Medicare, at 42 CFR 482.13, includes your right to make informed decisions about your care, described in the regulation as “being able to request or refuse treatment.” The same rule is careful about the other edge: that right is not a tool for demanding treatment that is medically unnecessary or inappropriate. You can decline a procedure. You can’t order one off a menu.

The roots run deep. The StatPearls clinical reference on the National Center for Biotechnology Information’s Bookshelf traces modern consent law back to a 1914 New York ruling holding that an adult of sound mind gets to determine what is done to his own body. Everything since has been an argument about the edges of that sentence.

This is the part people skip. Saying no is only protected if it’s informed, and informed means you got the same briefing you’d have gotten before saying yes.

That’s the standard checklist MedlinePlus describes for consent: the problem and why treatment is recommended, what the treatment involves, the risks and how likely they are, how likely it is to work, whether the decision is urgent or can wait and what happens if you wait, the other options, and effects that may surface later. A refusal built on that is informed refusal. A refusal built on a rumor is just a refusal.

So the first move when you want to decline something is not to argue. It’s to ask for the briefing you’d get if you were about to agree. What is this supposed to prevent, how likely is that thing without it, and what’s the realistic cost of skipping it this year?

A good clinician will then write the conversation down, and you should want them to. A note saying the risks of declining were explained and the patient understood them is not a trap. It’s a record that you were treated as a decision-maker rather than a problem. The hospital record rules already require signed consent forms for certain procedures to be filed in your chart, so documentation of the opposite decision follows the same logic. If a refusal form appears, read it the way you’d read a consent form: it documents a conversation, and if the conversation didn’t happen, say so before you sign.

Signing a refusal does not lock any doors. Changing your mind at any point is your right, and MedlinePlus says so on its patient rights page. Declining chemotherapy in March does not disqualify you from starting it in June.

Refusing one thing is not refusing care

This is the idea that changes how the rest of the article reads. A refusal is almost never all-or-nothing, and treating it that way is what turns a reasonable disagreement into a rupture.

Think about the pieces of a recommendation separately. There’s usually a drug or a procedure, a monitoring plan, a follow-up interval, and a set of warning signs that would change everything. You can decline the first and keep all the rest.

Three shapes a workable refusal can take:

Decline the treatment, keep the surveillance. “I don’t want to start the medication. I do want the blood work in six months, and I want to know what number would change your mind.” You’ve now turned a flat no into a monitored plan, which is far safer than walking away.

Defer, with a date. “Not this year. Let’s look at it again at my physical in October.” Put it in the chart. A deferral with a date is a decision. A deferral without one is drift.

Ask what would move you. “What would you need to see, in my labs or my symptoms, before you’d tell me this is no longer optional?” That question hands you a tripwire you can watch for, and it tells the clinician you’re engaged rather than dismissive.

There are real decisions where declining is an accepted outcome, not a deviation. Prostate cancer screening with the PSA blood test is the clearest example. In its 2018 recommendation, the US Preventive Services Task Force gave screening a C grade for men aged 55 to 69, meaning the choice is an individual one that should be made after discussing the potential benefits and harms with a clinician. The Task Force went further and said clinicians should not screen men who don’t express a preference for being screened. At 70 and older, the grade drops to D, and the panel advises against the PSA test. In that case, “no” is the default.

Tools exist to make these conversations less lopsided. A 2024 Cochrane review of patient decision aids, which pooled 209 randomized trials with more than 107,000 participants, found that decision aids produced large improvements in knowledge and in how accurately people understood their own risk, and made people less likely to sit passively while the clinician decided. One finding matters especially here: there was no difference in decision regret between people who used a decision aid and people who didn’t. Being more involved in the decision did not leave people second-guessing themselves more.

Four questions before you decline

Use these in the room, in this order.

  1. If I skip this, what’s the most likely thing that happens over the next year, and how likely is it?
  2. What would you watch instead, and how often?
  3. What symptom or result should make me call you right away?
  4. Is there a smaller version of this, a lower dose, a shorter course, a less invasive test, that gets most of the benefit?

Question four is the one people forget. A fair number of standoffs end because somebody finally mentions there was a middle option the whole time.

Where the right stops

The right to refuse is broad, not unlimited, and the limits are mostly set by your state rather than by Congress. MedlinePlus says it plainly: patient rights differ from state to state. Some are federal, like access to your records. Many aren’t, which means a confident claim about what the law requires is often only true in some states.

Emergencies. When waiting would put a patient in danger, MedlinePlus says the consent requirement falls away. The AMA’s informed consent opinion describes the same exception: if the decision is urgent, you can’t take part, and no surrogate is available, treatment can begin without prior consent. This is a situation where nobody is in a position to refuse, not a refusal being overridden.

Capacity. The right belongs to people who can exercise it. Three things make up that ability in the StatPearls account: you absorb and hold on to the relevant information, you weigh the options, and you say which one you want. A diagnosis by itself does not erase them. Someone with cognitive impairment, a mental illness or a severe illness keeps the right to decide, provided those three abilities hold. A decision a clinician disagrees with is not evidence that capacity is missing. When capacity is genuinely in question, StatPearls notes that a psychiatrist may be asked to evaluate it.

Decisions for minors. A parent or guardian agrees to a child’s care, and StatPearls calls that informed permission rather than consent, which means the power to say no sits with the adult too. A legally emancipated minor can consent without them. State laws differ substantially, including on which decisions older teenagers can make alone.

Public health and court orders. Some refusals run into state law rather than medical ethics, in areas like communicable disease control, school immunization requirements and court-ordered evaluation or treatment. These are state statutes, they are not uniform, and this site isn’t going to tell you what yours says. If you’re in that territory, that’s a question for a lawyer licensed in your state, not a health article.

When you can’t speak. The way to make a refusal stick in advance is to write it down. MedlinePlus describes two standard documents: a living will, which specifies which treatments you want or refuse if you’re dying or permanently unconscious, and a durable power of attorney for health care, which names a person to decide for you. The federal hospital rule at 42 CFR 482.13 says patients have the right to formulate advance directives, and that staff and practitioners must follow them. A refusal you never recorded is a refusal somebody else gets to interpret.

The risk this site is not going to soften

Refusing a specific drug is one thing. Refusing effective treatment for a serious disease in favor of something unproven is a different thing, and the evidence on it is not ambiguous.

There’s a distinction from the National Center for Complementary and Integrative Health that’s worth borrowing here. Using a non-mainstream approach alongside conventional medicine is complementary. Using it instead of conventional medicine is alternative. NCCIH notes that genuinely alternative use, as a replacement, is rare. It’s also the version where the outcome data turns grim.

Two studies published in 2018 looked at this directly in US cancer registry data. One appeared in the Journal of the National Cancer Institute. Its authors tracked 281 people with non-metastatic breast, prostate, lung or colorectal cancer who turned down surgery, radiation, chemotherapy and hormone therapy and used alternative medicine alone. Matched against patients who received conventional treatment, they had about two and a half times the risk of death. For breast cancer specifically the hazard ratio was 5.68.

The second study, in JAMA Oncology, drew on the National Cancer Database and is the more interesting one, because it separates two things people routinely confuse. Patients who used complementary medicine alongside conventional treatment had worse five-year survival, 82.2% versus 86.6%. But when the researchers accounted for treatment refusal and delay, the association largely disappeared. The complementary therapy itself wasn’t what raised the risk. Refusing the conventional treatment was, and people using complementary medicine refused far more often: 34.1% declined chemotherapy compared with 3.2% of the matched group, and 53.0% declined radiation compared with 2.3%.

Read that carefully, because it cuts both ways. Adding acupuncture or a mind-body practice to your cancer care was not the danger in that data. Letting it displace the treatment was. The same logic applies far outside oncology, to insulin, to blood pressure medication, to antibiotics for a real bacterial infection.

If what you’re declining is treatment for a condition that kills people, get a second opinion before you decline rather than after. And be honest with yourself about the question underneath the refusal: are you rejecting this specific treatment, or are you reacting to how you were treated by the person who recommended it? Those need different fixes.

Walking out of a hospital early

Leaving against medical advice is a refusal too, and it’s the one with the best American data.

A study in Annals of Epidemiology, using the Nationwide Readmissions Database from 2010 through 2017, reported that 1% to 2% of hospitalized US patients each year leave against medical advice. Their 30-day readmission rate was 25.6%, and after adjustment their odds of being readmitted ran roughly double to quadruple those of patients discharged normally.

That number is not an argument that you must stay. Sometimes people leave for reasons a hospital can’t fix, like a child at home or a job that won’t hold. It’s an argument for not leaving empty-handed.

If you’re going, ask for four things on the way out: the prescriptions you’d have been sent home with anyway, a follow-up appointment, written return precautions naming the specific symptoms that mean come back now, and a copy of what’s been done so far. Staff can usually provide all of it. A planned early exit and a storm-out produce very different outcomes.

When a clinician takes it badly

Most won’t. Some will, and you should know what you’re entitled to.

The AMA’s patient rights opinion lists courtesy and respect, answers to your questions, having your decisions respected, a copy of your records, and a second opinion. It also covers what happens if a physician decides to stop treating you: under the AMA’s opinion on terminating a patient-physician relationship, physicians have an obligation to support continuity of care, which means notifying you far enough ahead that you can find another physician and helping transfer your care. A doctor can decline to keep managing a case. A doctor is not supposed to drop you without notice because you said no to something.

If things go wrong inside a hospital, there’s a formal channel most patients never use. The same federal patient rights rule that gives you the right to request or refuse treatment also requires hospitals to run a grievance process with set review timeframes and written decisions. MedlinePlus points to hospital patient advocates, state long-term care ombudsman offices for long-term care problems, state health departments, and the HHS Office for Civil Rights for discrimination or privacy complaints.

One practical move that costs nothing: ask that your concern and the clinician’s response both go in the visit note. People behave differently when the conversation is being written down, and you end up with a record you can point to later.

And sometimes a bad reaction is information. If a clinician cannot tolerate being told no about one medication, that tells you something about how the next ten years of decisions will go.

The single most useful thing here

Separate the refusal from the relationship.

“No” to one treatment is a medical decision. Walking away from monitoring, follow-up and the person who knows your history is a different decision, and it’s usually the one that hurts people. Decline the drug and keep the appointment. Defer the procedure and set the review date. Ask what would change the recommendation, then watch for it.

If you’re about to decline something for a condition that can kill you, do two things first: get a second opinion, and ask the clinician recommending it to tell you, in a number, what happens over the next year if you don’t. You may still say no. You’ll be saying it with your eyes open, which is the whole point of the right.

Sources

  1. MedlinePlus, National Library of Medicine. “Informed consent – adults.” medlineplus.gov
  2. MedlinePlus, National Library of Medicine. “Patient Rights.” medlineplus.gov
  3. MedlinePlus, National Library of Medicine. “Advance Directives.” medlineplus.gov
  4. StatPearls, NCBI Bookshelf, National Library of Medicine. “Informed Consent.” ncbi.nlm.nih.gov
  5. American Medical Association. “Patient Rights: Code of Medical Ethics Opinion 1.1.3.” code-medical-ethics.ama-assn.org
  6. American Medical Association. “Informed Consent: Code of Medical Ethics Opinion 2.1.1.” code-medical-ethics.ama-assn.org
  7. American Medical Association. “Terminating a Patient-Physician Relationship: Code of Medical Ethics Opinion 1.1.5.” code-medical-ethics.ama-assn.org
  8. Electronic Code of Federal Regulations. “42 CFR 482.13, Condition of Participation: Patient's Rights.” ecfr.gov
  9. US Preventive Services Task Force. “Prostate Cancer: Screening.” uspreventiveservicestaskforce.org
  10. National Center for Complementary and Integrative Health. “Complementary, Alternative, or Integrative Health: What's In a Name?.” nccih.nih.gov
  11. PubMed. “Johnson SB et al. "Use of Alternative Medicine for Cancer and Its Impact on Survival." Journal of the National Cancer Institute, 2018. PubMed. DOI: 10.1093/jnci/djx145.” doi.org
  12. PubMed. “Johnson SB et al. "Complementary Medicine, Refusal of Conventional Cancer Therapy, and Survival Among Patients With Curable Cancers." JAMA Oncology, 2018. PubMed. DOI: 10.1001/jamaoncol.2018.2487.” doi.org
  13. PubMed. “Spooner KK et al. "Increased risk of 30-day hospital readmission among patients discharged against medical advice: a nationwide analysis." Annals of Epidemiology, 2020. PubMed. DOI: 10.1016/j.annepidem.2020.07.021.” doi.org
  14. PubMed. “Stacey D et al. "Decision aids for people facing health treatment or screening decisions." Cochrane Database of Systematic Reviews, 2024. PubMed. DOI: 10.1002/14651858.CD001431.pub6.” doi.org