Share this article on:

Discharging Patients Against Medical Advice

Roberts, James R. MD

In Focus: Part 1 in a Series

Standard AMA forms lack total liability protection, and the best course is to avoid AMA discharges whenever possible

A compendium of Dr. James Roberts' InFocus columns is available in book form. The 302-page volume, InFocus: Roberts' Practical Guide to Common Medical Emergencies, is available from Lippincott Williams & Wilkins for $59.95 by calling (800)638–3030.

Author Credentials and Financial Disclosure: James R. Roberts, MD, is the Chairman of the Department of Emergency Medicine and the Director of the Division of Toxicology at Mercy Health Systems, and a Professor of Emergency Medicine and Toxicology at the Drexel University College of Medicine, both in Philadelphia, PA. Dr. Roberts has disclosed that he has no significant relationships with or financial interests in any commercial companies that pertain to this educational activity. Wolters Kluwer Health has identified and resolved all faculty conflicts of interest regarding this educational activity.

Release Date: June 2005



Learning Objectives: After reading this article, the physician should be able to:

  1. Discuss the issues involved in AMA discharges.
  2. Describe the proper use of the AMA form.
  3. Explain the issues that should be addressed before allowing a patient to leave the ED.

Emergency physicians have enough to worry about when it comes to diagnosing and treating ED patients. We all have been in the situation where, after expending considerable time, effort, and resources, the patient suddenly announces that he really does not want to take our advice, and has decided to leave the hospital. It can occur before all the tests are back, but it's more problematic when the decision to bolt is proclaimed after all the paperwork and consultation are in place for an admission for documented pathology. This is a frustrating dilemma, and one that is fraught with a plethora of potential and real complications for the patient and physician.

During an exhausting ED shift and after extending tremendous effort on the patient's behalf, it is only human nature for the physicians and nurses to become angry, insulted, and annoyed at an unexpected AMA announcement. The staff may become as unreasonable as the fleeing patient, hastily reacting to an unreasonable demand or creating a very hostile and unfriendly situation for both parties. It's also very tempting to simply comply with the patient's wishes and reach for the AMA form. We all try to discuss the issues and dangers of leaving against medical advice rationally, but once a patient makes up his mind to leave, it may be difficult to dissuade him from expediting this poor decision.

A normal reaction from an overworked and frustrated medical staff is to ask the patient to sign a vague and uninformative AMA form written 10 years earlier, and allow the patient to fend for himself, branding him ignorant or ungrateful. We all have certain tricks, cajoling scenarios, negotiating tactics, and bargaining chips to play, but in the end, when it's impossible to change the patient's mind, it's time for a critical decision: allow the egress or restrain/sedate the patient. A “good riddance” mentality may allow the neophyte clinician to reach for the AMA form to settle the issue quickly. However, to the sagacious veteran physician, things are much more complicated than they initially seem. My advice is to attempt to keep the patient in the system at all costs, most certainly if he has a proven medical problem.

As the plaintiff lawyer always says in the courtroom, “Doctor, isn't it true that you allowed this patient to sign his life away?”

Everyone loses in an AMA discharge. Maybe the patient is just being difficult and is so unreasonable that even Herculean efforts will not produce even a modicum of success to keep him in house. After all, there is no law against being stupid, rude, or foolish. But the AMA discharge can be a failure to communicate or a failure of the system. If an individual is drunk, drugged, or mentally incompetent, no one would allow him or his family to orchestrate leaving, but in questionable cases and especially if the patient seems perfectly capable of making a bad decision, our hands may be tied.

This month's column begins a series of discussions on patients who want to leave the ED against medical advice. I will discuss the value of the AMA form, what happens to such patients after they leave, strategies to keep patients from being their own worst enemies, and how to protect the physician against the omnipresent legal ramifications. I begin by investigating the true value of the standard AMA form.



Many patients who leave AMA have serious problems, some of whom will have a bad outcome after their ED experience. A properly executed AMA form (note the caveat) will quell most reprisals against the doctor or hospital. There is still some self-accountability left in the land, but contrary to popular belief, the AMA form may not protect the hospital or physician totally when there is a bad outcome. It's difficult for the family or public to believe that something more could not have been done to sidetrack the irrational behavior of someone who is obviously ill or in need of hospitalization or treatment. Unfortunately, merely having him sign a piece of paper is not blanket legal protection for the emergency physician in all cases. In many cases, an AMA form is totally worthless, and a poorly written one is likely worse than none at all.

Does Identifying a Discharge as Against Medical Advice Confer Legal Protection?, P Devitt, et al, J Family Practice, 2000;49(3):224

The authors of this publication from the Department of Psychiatry at the State University of New York note that AMA discharge from general hospitals range between one in 65 and one in 120 ED visits. Patients most likely to fall into the category of AMA discharges are younger, male, and seen in the ED. Many have been hospitalized frequently and live alone. Some studies note a higher rate of AMA discharge in the elderly. Occasionally psychiatric consultations are obtained to assess mental competency when the medical staff perceives imminent or serious danger to the patient.

Other factors involved in the decision to leave AMA include patient anger, fear, and psychiatric illness. Although patients discharged AMA may have minimal or inconsequential medical problems, many do have an exacerbation of an underlying serious illness, and leaving the hospital will result in adverse consequences, including exacerbation of an illness, death, or injury to the patient or others. Even in the absence of negligence, an adverse medical consequence of an AMA discharge may lead to a malpractice suit.

If a patient rejects advice on admission or other medical interventions, an AMA form is thought by many physicians to confer some sort of blanket legal protection. The literature is vague on this issue. There is no well promulgated standard of care. The authors attempted to determine whether an AMA discharge is protective, partially protective, or offers no protection at all. Using a computer search of the literature, the authors found relevant legal documents and addressed particular cases.

As of the year 2000, the authors found only eight civil cases in which an AMA discharge was a significant issue in a medical malpractice case. It is unclear how many times an AMA discharge actually is problematic.

A common physician and hospital defense tactic for an adverse outcome is to claim that the patient contributed to the ultimate complications by discharging himself from the hospital against medical advice. This is called contributory negligence. As one example, the authors briefly review the case of a woman who after many years of chronic abdominal pain and four weeks of hospitalization discharged herself AMA before a definitive diagnosis had been made. She eventually died from ischemic bowel disease. To claim contributory negligence, one would have to show that the patient acted unreasonably. In a somewhat confusing explanation, a jury found that the patient did not act unreasonably when she discharged herself from the hospital. For some bizarre reason, the court found that because she had been ill for a number of years, had been in and out of hospitals, had a number of tests and was frustrated, it was not unreasonable for her to leave the hospital against advice. Therefore, the medical system was at fault! (I still cannot believe this case.)

In another case, a reduction in damages was awarded, but the hospital was still found liable when a patient was discharged AMA, concluding that the patient's actions merely increased the extent of injury, but did not cause it. In a third case, even though the patient left AMA, missed two scheduled appointments, failed to take his medications, and injected himself with street drugs, it was still somehow the hospital's responsibility for adverse outcomes.

Although an AMA form is a common form, there is little recourse for physicians when a patient refuses to sign such forms. In such cases, it's up to the clinician to construct his own version of the incident.

As a general rule, patients are admitted voluntarily to a hospital, and an AMA discharge is merely withdrawal of original consent for treatment. Although competent adults possess the ability to make this decision, it is the responsibility of physicians and hospitals to ensure that the patient is fully informed of the risks and alternatives to treatment, possesses mental competency to make a reasonable decision based on available information, and does not meet standard criteria for involuntary psychiatric hospitalization. Unless all three of these issues are carefully documented, physicians and hospitals may not be legally protected in an adverse outcome. Clearly, the term “against medical advice” is not by itself protective. The authors stress that physicians should ensure and thoroughly document that a patient is mentally competent, fully informed, and does not meet the criteria for involuntary hospitalization before he is allowed to make his own decisions about care.

In summary, the authors could find no consistent evidence that a patient discharged AMA would not be able to sue a hospital or physician successfully for a subsequent adverse outcome. Although careful and thorough documentation is the first line of defense, numerous intricacies and details are often not in place. Physicians must be able to show that patients were fully informed about the specific risks of leaving the hospital, and that there were alternative venues of care. Failure to attempt to arrange follow-up or alternative care or to provide outpatient care can be considered malpractice. Essentially, the authors believe that requiring a patient to sign a waiver releasing the hospital from all responsibility following an AMA discharge would be considered legally worthless. Although the AMA form may provide partial protection, it is certainly not a guarantee of legal immunity.



Comment: I find this article difficult to decipher, but I would agree with its conclusions, however unsettling and counterintuitive. It is essentially a medical myth that having a patient sign an AMA form protects the physician or hospital from successful litigation when the patient or someone else suffers an adverse outcome. One common scenario is when a discharged patient injures someone else, such as by driving a car while intoxicated or under the influence of a medication given in the ED. Other common cases are psychiatric patients, where the psychiatric illness is underestimated or unappreciated.

I could find little in the medical literature that offers the physician particularly stellar advice about AMA discharge. In essence, there is no standard of care, but there is a lot of naivet and hubris on the part of physicians. In my experience, EPs do a poor job of writing a good AMA form. Over the years I have developed some criteria, caveats, and common sense rules about AMA discharges. I have not found these particularly highlighted in the medical literature, but offer them to the prescient clinician who wants to stay out of court or win a case if he finds himself in such a situation. My first goal is to have the clinician realize that the AMA form is not blanket protection from a lawsuit or a successful claim of negligence. My second is to convince physicians to try to avoid AMA discharges and fight tooth-and-nail for the patient who has no clue how sick he really is or what dangers might befall him on the outside.

In my opinion, the standard AMA form that is hurriedly signed by the patient as he angrily leaves the hospital is essentially a worthless document. Few physicians or nurses have read it, and it is too general for most cases. Try reading the one that your hospital uses, and then show it to a nonmedical family member for an opinion. Generally it's an entire paragraph consisting of one sentence, with print too small to be seen without a magnifying glass.

Likely you will not read it to the exiting individual, and certainly he will not read it before he signs it. Sometimes it contains some unintelligible legal mumbo-jumbo written by a hospital lawyer 10 years before, claiming that you and the hospital are not responsible for anything if the patient leaves or refuses a treatment plan. Don't believe it! The parking lot may say it is not responsible for lost articles in your car under its care, but gross negligence negates any disclaimer in the real world. Your hospital probably has a one-size-fits-all document, and we all know that just doesn't make it. I believe every case is different, and every patient encounter should be individualized. Therefore, everyone should have his own specific AMA form. With a little bit of my paranoia, I have found that the template chart by T-System (modified by me for our hospital) meets most of the criteria for a personalized document. At least this one addresses specific issues of cognition, comprehension, risks, follow-up, and other details that can be documented for individual situations.

Simply stating that the mental status is intact or that the patient appears to understand are dangerous generalizations. Better yet, write your own AMA note on every patient to fit the specific scenario, and make sure you include the required documentation. “Oriented times 3” is not a complete mental status examination, but unfortunately that is all that may appear in a written record. I also like to have the nurse write a paragraph in the nurse's note about the circumstances. There should be some mention about clear speech, normal gait, efforts to resolve conflicts, and a nonhostile interaction that allows the patient to save face if he thinks things over.

Because many patients who decide to leave AMA have psychiatric issues or are drug users, this subgroup requires special attention. It is very difficult to convey on paper the fact that a patient with a positive drug screen, an elevated alcohol level, or a complex past history of psychiatric disease and on multiple medications completely understands complex medical terminology or consequences of leaving the hospital.

Functional illiteracy of ED patients is omnipresent and well known, so any AMA form should contain simple, straightforward wording. (Ann Emerg Med 1988;17[2]:124; Ann Emerg Med 1993;22[3]:573). The statement, “Dr. Roberts told me I could die if I go home,” works well for me. In some cases, I write this directly on the ED chart, and have the patient and preferably also a relative sign the note in their own (legible) handwriting. This is in addition to the AMA form. Also, make sure the AMA form is mentioned somewhere in the ED chart because extraneous pieces of paper, such as discharge instructions, tend to get lost in medical records. I have had patients deny the entire discussion, and claim they never received any information or paperwork. Of course, there is always the patient who will not sign the form. That is not a problem; simply fill it out for him, and note the refusal to sign.

Psychiatric patients are particularly problematic. Often the Psych Crisis Department makes the final call on this one, but the ED usually makes decisions about concomitant medical issues on the psychiatric patients. Psychiatric patients are impulsive, difficult to deal with, argumentative, and often know how to manipulate the medical system to their advantage. However, psychiatric patients do kill themselves, and they also kill and injure other people. Bipolar patients have an extraordinarily high incidence of successful suicide compared with other conditions. If a bipolar patient discharged AMA kills someone within a few hours of ED discharge or commits suicide soon after leaving your ED, it would be very difficult to convince anyone that he had a normal mental status or was not suicidal or homicidal when you saw him.

Although the general public expects a psychiatric consultation under such circumstances, I have not found that route to be particularly helpful. In fact, I believe that our psychiatric colleagues are even less tuned in to the ED patient and have little ability to prognosticate medical illness. Also, I don't know how they could be more astute in determining subsequent behavior than an emergency physician who likely has spent more time with the patient and saw him on arrival, often before the AMA decision was hatched. No one can predict suicide, and many cannot be prevented by any intervention. Having a psychiatrist clear a patient for an AMA discharge may save the EP from some portion of the litigation, but it is not the ultimate answer.

In addition to a psychiatrist, it's also helpful to involve other individuals, a proactive tactic that clearly states that you are a patient advocate and have done everything possible to keep this patient from making the wrong decision. It's a good idea to offer transfer to another facility (usually impossible, but you tried). Also, offer to call the patient's physician or to speak to a family friend or relative. With the HIPAA rules, some of these things may be questionable, but I have no qualms about calling the emergency contact number on the patient's chart before I let him go. If the patient refuses to give this information, document it. Also, going to the waiting room to see if there is a friend or relative may be productive. Having a drunk common law spouse complicate matters further is not helpful, but often a caring son, daughter, or spouse can convince a patient to stay when the physician has been unsuccessful. If you talk to someone such as the patient's doctor or a relative on the phone, hand that phone to the patient and document that you did so.

The authors of this study suggest that one make a genuine attempt at providing follow-up or alternative care for the patient who is leaving AMA. Failure to do so can be interpreted as a breech of duty of care. That's really difficult in my experience, especially when you just tried to provide a great follow-up — hospitalization. Just because a patient with chest pain and unstable angina or MI wants to leave the hospital, that's no reason to withhold nitroglycerine, beta blockers, or aspirin. It's usually pointless, but it's easy enough to provide the patient with a cardiologist's phone number. Similarly, providing the patient with a prescription for antibiotics for an infection that should be treated in the hospital certainly could be considered proactive and patient friendly. If the patient promises to call his personal physician as soon as he gets home or go to the office or another ED, document that on the chart. It's also a good idea to inform the primary physician when his patient invokes the AMA option.

It also should be made crystal clear on the chart that the patient has the ability to return to the ED at anytime, should he change his mind or should he get worse. It may be difficult to have an amicable separation under AMA circumstances, but it's best to maintain an open and supportive environment. It's only human nature for a disgruntled patient to feel ashamed, embarrassed, or too angry to return to the hospital that provided the primary evaluation. If he leaves on friendly terms, he may be more likely to return when he rethinks his decision with a pain, infection, or shortness of breath that persists or becomes worse. The door to the ED should always be open, and the patient should not feel that the ED will hold anything against him simply because he chose to leave.

Table. Cavea

Table. Cavea

Many patients may leave the ED because they fear a large medical bill or because they have no insurance. The ability to pay should never be an issue with any emergency encounter. It should be explicitly stated in the medical record that patients are not discriminated against because they do not have the right insurance or cash. That may be a luxury of the private physician or consultant, but the ED must make it crystal clear that money is not a barrier to care. I prefer to address it up front with a pre-emptive strike, and state in no uncertain terms that the ability to pay for medical care has not clouded my decision, the patient's decision, nor access to the system.

Finally, restraining patients against their will is omnipresent. Most physicians will say that competent patients have the ability to make a poor decision that may ultimately adversely affect their health in a significant manner. This is true, but I believe this is only partly true. Many cases are borderline, complex, and fraught with a variety of subtleties and nuances that are impossible to convey in the medical record. Two years later, many different interpretations and alternative versions of the encounter come to light, and it is difficult to rely on recall or a less than ideal medical record. Many of my colleagues disagree, but I will sedate or restrain a patient who has a serious medical problem if there is a way to justify it, and there almost always is. That's probably not standard of care, but this is a gray zone. I always ask myself the question, What would I rather defend: restraining a patient, sedating a patient against his will (AKA assault), or three years later try to defend sending home a patient with a medical condition that I thought could kill him or leave him significantly impaired? Despite concerns of unlawful restraint, I have yet to find this to be anything more than a philosophical issue that goes no further than a nasty letter to an administrator.

Although I know it has been tried, it's difficult to sue a doctor for saving a life or limb. One can always rationalize the use of Ativan or morphine, even Haldol, as a method to reduce anxiety or pain, and therefore allow the patient to think more clearly or make a more rational decision by relieving stressful factors. Often a little bit of sedation goes a long way in resolving the issues. At least it buys the physician time to garner reinforcements, rethink the strategy, or obtain additional data. While patients have the ability to make their own decisions, often times they need a little help from readily available medication, and even a gentle hand on the shoulder to keep them in the ED.

You also must make certain that the pulse oximetry and glucose are normal, the alcohol level and drug screen are acceptable, and that patients are not leaving AMA because you refuse to treat their pain or other symptoms. Don't let your withholding a few Percocet make a patient mad enough to walk out before you are ready to discharge him. Let him know that dinner is on the way, you are expediting the bed, process, or test, and that you won't force an NG tube, central line, or Foley catheter in borderline cases.

Although every patient does not need an alcohol level or drug screen prior to an AMA discharge, you must be certain that the nurse has not put down “AOB,” that you have not documented needle tracks, or that a well meaning triage nurse has not noted bizarre or irrational behavior or suicidal thoughts prior to your entering the room. Always read the nursing notes and address issues that you cannot confirm or are no longer present!

Unfortunately, there are many areas in medicine today where the physician just can't win. AMA discharges with a bad outcome are high on the list. It's not fair, but it is a reality. If a patient leaves AMA, it is often the physician's failure to establish a good initial patient rapport, the inability to convey the seriousness of a disease state, or inability to understand the patient's viewpoint and address minor issues that can be readily corrected. It doesn't make sense that the patient in the throws of a myocardial infarction wants to go home to feed his cat, but instead of reaching for the AMA form when this excuse is given, offer to call a neighbor and let him feed the feline. Some might interpret the cat excuse as irrational behavior and not allow discharge. Others could interpret leaving the ED in the midst of an aortic dissection as a suicide gesture, and likewise nix the egress.

You also can involve social services to help with child care at home, a visiting nurse for a sick relative, or some such careful, conscientious, and prudent interaction. While most times the AMA form will be a sanctuary for the clinician/hospital, there is no AMA form on earth that will convince a wife that you could not have done something to keep her husband in the hospital when everyone in the room knew he was having a heart attack. Legally he could make that decision, but you don't need to facilitate it.

I have been personally involved in attempts to defend two physicians who were at the wrong end of an AMA discharge. I lost both arguments with an out-of-court settlement against the physician. In one case, a patient with vague chest pain was allowed to sign out AMA, and died on the way home from the hospital of VF and an acute MI. The patient had signed an AMA discharge form, and the physician had written a relatively lengthy note describing a valiant attempt to keep him in the ED. Unfortunately, he never spoke to the family who picked him up in the ED waiting room, and did not time the note. It was interpreted that this was written when the patient came back coding in the ambulance. Note to self: Always time your AMA discharge notes.

In a second case, the patient was allowed to sign out AMA after he was seen in the ED for a minor complaint. The nurse had noted alcohol on his breath and some agitation. The patient became annoying and wanted to leave, his minor complaint properly addressed. The standard AMA form was signed without the personal involvement of the physician or strict attention to detail, and when the patient ran off the road and hit a tree 20 minutes later, his ethanol level was almost 300 mg/dl. Although an alcohol level may not always be required, it certainly would have been good form for the emergency physician to assess the mental status, gait, and cognitive function carefully prior to allowing the nurse to sign out this patient who was becoming problematic and rude. Second note to self: The physician always does the AMA discharge, not the nurse.

Everyone loses in AMA discharges. Try to avoid them when you can. While a pristine, squeaky clean AMA form is better than none at all, you are fooling yourself if you think it will always completely protect you from litigation, or that it is the simple answer to a problematic patient encounter. As the plaintiff's lawyer always says in the courtroom, “Doctor, isn't it true that you allowed this patient to sign his life away?”

Copyright © 2005 Wolters Kluwer Health, Inc. All rights reserved.