Feature Science in the News Stereolab PubMeditation Culture Dish Sportsology Laberrations The Mad Gastronomist Mass Spectator LD50 Home Home

Move Aside Doctor; It’s Time for Me to Go

By Robert Roose, MD MPH | 4.30.07

It was a typical Thursday night on the inpatient family medicine floor. There were twenty-one patients and two empty beds, not counting the folding screen propped up in the hallway that too often doubled as a patient room. On one end of the floor, in Room 254, Mr. Sanchez was being treated for pneumonia. In Room 256, Mr. Trong, suffering from a massive brain tumor, was being medicated for agitation. Down the hallway, in Room 263, Mrs. Holloway had just spiked a fever and was feeling faint. The intern on-call was busy writing up an admission note. The nurses were in the lounge, finishing up their late-night meal of beef patties and curried goat. Nearly everyone was asking for a sleeping pill. And rolling off the elevator into Room 259 was Frank Bryant, the next patient to be admitted from the overflowing emergency room.

Mr. Bryant was a forty-eight year old man with HIV/AIDS who was having difficulty breathing, and had recently been hospitalized for the same complaint. In fact, that night, he was being admitted for a process that started three weeks earlier when he presented to the same ER with leg swelling and shortness of breath. There he was found to have evidence of cardiac muscle damage, likely from severe heart failure, and was admitted to a cardiac monitoring unit. His physicians prescribed medications to help reduce the amount of fluid he was accumulating and ordered a slew of blood tests and an echocardiogram. However, Mr. Bryant was not an easy-going sort of guy. Over the first few days of his hospitalization, he refused pretty much everything at some point, including blood draws, imaging tests, physical examinations, and answering questions. When he took the medications that were ordered, he improved. But with all of his refusals, he was only receiving occasional treatment, and so each bit of progress was countered by at least as much failure.

By the end of the first week, the frustration of his doctor, a well-regarded and even-tempered hospitalist, was increasingly evident in her daily “progress” notes. She documented spending unusually large amounts of time with him, trying to understand his resistance to treatment in order to establish some sort of working relationship, much less a therapeutic alliance. Yet Mr. Bryant remained difficult to care for. He requested “specialists” several times, but when they came by to evaluate him he would not talk to them. Although he did finally agree to have an echocardiogram, which confirmed global, severe heart dysfunction (his heart pumped at about one-third the strength of a normal heart), he refused a cardiac stress test to better assess his coronary arteries. He continued to take his heart failure medications only sporadically. He declined treatment for HIV/AIDS. Whereas most patients with his condition would have shown significant improvement in three or four days, after one week Mr. Bryant seemed to only be getting worse.

At that point a psychiatry consult was called. And after a long discussion with him, the attending psychiatrist decided that although he exhibited clear paranoia, he was not delusional. He understood the risks and benefits of his decisions and had the mental capacity to refuse as little or as much treatment as he wanted. As long as the refusals were not regarding an imminent life-or-death decision, he could not be forced to oblige, and he did not meet criteria for an inpatient psychiatric hospitalization.

Later that night, the nurse was called because he was coughing up blood. She took his blood pressure and pulse, which were both normal, and called the doctors, who immediately thought infection. Typical bacterial pneumonia can often cause blood-tinged sputum; but since he had been in the hospital for so long, he was now also at risk for infections caused by Methicillin-resistant staphylococcus aureus (MRSA) or Pseudomonas aeruginosa; and because he had HIV/AIDS, it could also be pneumocystis carinii (PCP) or tuberculosis, all of which were potentially life-threatening if not treated. They immediately started him on intravenous fluids and two powerful antibiotics. Of course, although the situation had changed, Mr. Bryant’s response had not. He would not let the staff draw blood cultures to look for infection. He would not stay hooked up to the fluids. And he had decided that he only wanted one of the antibiotics, and once a day instead of three.

So it went for two more weeks. One day Mr. Bryant would take some antibiotics, the next he would not. At times he would let his doctors evaluate him, and other times he would refuse. Cardiologists, gastroenterologists, pulmonologists, internists, psychiatrists, nurses, social workers, and residents all talked with him, and they all failed to make a connection with him such that he would adhere to a reasonable treatment plan. He may have been feeling different than when he came in, but it was hard to say whether it was any better. Instead of florid heart failure, he now had severe pneumonia, and without a consistent antibiotic regimen, it was probable that the spotty treatment was only making the infection worse, by baiting antibiotic resistance. Nevertheless, three weeks to the day after his admission, Mr. Bryant decided that he wanted to leave – and so he did.

Not every patient that is admitted to the hospital is properly discharged by his or her physician. Some, unfortunately (or quite naturally), do not recover and die. Others, like Mr. Bryant, decide it is time to go despite their physician’s cautions and sign themselves out “against medical advice.” It is a decision that often smacks of trouble for the patient and smells like a “cover your ass” move on behalf of the physician. But is it truly either?

Over the past half-century the concept of leaving against medical advice (AMA) has changed substantially. In the 1950s, psychiatric patients were considered discharged AMA only if they managed to escape from the hospital and did not return within a certain amount of time (something now called elopement). In the 1960s and 70s, advocacy efforts and legislative changes gave patients increasingly more control over their medical care, including the right to sign out AMA from a hospital. Then over the next few decades, with community support, the nature of AMA discharges started to change. The stigmatization of hospitalization waned and patients began to feel more empowered (both good things, of course). However, with this newfound hope came such a rapid rise in AMA discharges that they quickly became routine. Now, they are so common that physicians or patients rarely bat an eye. They occur in all settings, from the emergency room to the intensive care unit, and in all types of patients, from those with chest pain to psychosis. Physicians and nurses are well versed in how to deal with them. Hospital wards have ready-made consent forms and protocols just for this purpose. And patients in the know simply refer to it in slang as “signing out.”

On average about 1 to 5 percent of hospital discharges from medical services in the United States are against medical advice. However, the actual rates depend a lot on the type of patients. In one study of patients with HIV infection, many of whom were injection drug users, the rate was about 13 percent overall. In other studies, patients without health insurance or those who were eligible for Medicaid (state-sponsored insurance for low-income individuals) had twice the likelihood of being discharged AMA. Patients who leave AMA are generally younger, male, substance users, do not have a personal physician, and have more severe symptoms at discharge than those properly discharged by their physicians. They also tend to be angry and mistrustful upon discharge, and are more frequently hospitalized. In other words, they are a lot like Mr. Bryant.

When Mr. Bryant rolled up to the floor that night, it was less than twenty-four hours after he left the hospital AMA. Contrary to what he had said the day before, it did not seem like he was “better off at home.” He was having difficulty breathing and the oxygen levels in his blood were low. His legs were also quite swollen and he admitted not taking any medication while he was out of the hospital. Although he did not allow any blood tests in the emergency room, he did have a chest x-ray, which asserted what everyone, but perhaps Mr. Bryant himself, had feared: a profound worsening of his condition. The x-ray showed two large, cavitary lesions in his lungs that were likely either empyemas (walled-off collections of pus and bacteria, and the product of his poorly treated pneumonia) or dead tissue as a result of thrown blood clots. Either way, Mr. Bryant’s situation did not look good – and leaving the hospital certainly had not helped him.

Unfortunately, such is often the case. Patients who leave AMA are across the board more likely to be readmitted to the hospital, especially in the immediate post-hospitalization period. At St. Michael’s Hospital in Toronto, which serves a large low-income and sometimes homeless population, almost one-quarter of the patients who left AMA were readmitted within 15 days, compared to three percent for other discharges. In another study, about one-third of patients who left AMA were readmitted within 30 days. These patients were also more likely to be readmitted for the same diagnosisand stay in the hospital longer. Of course, on the flip side, some patients that leave AMA might be making good decisions. There is mounting data to suggest that longer hospital stays are not necessarily better; it is possible that some patients may be able to assess improvement before their physicians can; and in some cases, there may exist other life circumstances that supercede another day (or two or three) in the hospital.

In Mr. Bryant’s case, however, that seemed unlikely. He just did not have a good track record. In a highly supervised setting, he was not consistent with his medications. He was uncooperative with staff. Despite the psychiatrist’s evaluation, he just did not seem to completely grasp the gravity of his situation. And so from the moment he signed out, unofficial odds were that he would suffer some adverse consequences, including worsening of his illness or death.

So what should a physician do in this situation? We do not have a good way to predict which patients are more likely to be readmitted. There is no data on differences in death rates or disease-specific consequences for patients that leave AMA. And, first and foremost, physicians have the obligation to respect a patient’s autonomy. At the end of the day, even in the face of serious risks, it remains their right to choose. But physicians also don’t want to be held responsible when that patient with chest pain who leaves the emergency room before his blood tests come back ends up having a fatal heart attack. So a common slogan in hospital medicine has emerged: “If they want to leave, fine, but make sure they sign out AMA.”

The rationale behind this statement is that having a patient sign out against medical advice confers some sort of legal safeguard to the physician. However, in reality, it doesn’t hold much water. Patients that leave AMA can still bring malpractice claims against their physician. In fact, some of the factors that may contribute to a patient leaving against medical advice, such as anger and psychosis, actually increase the chance of legal action. Since admission to a hospital is voluntary, signing out AMA is merely a withdrawal of the original consent to be admitted. Just as any competent adult can accept admission, anyone can decide to leave AMA. It is the medical staff’s responsibility to ensure, as Mr. Bryant’s physicians did, that the patient is informed of the risks and alternatives to leaving, that the patient is competent to make that decision, and the patient does not meet criteria for involuntary psychiatric hospitalization. Most hospital consent forms have clauses in them to alert the physician to these important safeguards, because if these steps are not taken, the hospital may be particularly vulnerable in the event of a bad outcome. Yet even if appropriate measures are taken, having a patient sign out AMA does not always offer protection. Courts have decided that in some cases leaving AMA is a reasonable decision, and if so, it provides no protection for the physician.

Although a malpractice suit was not anyone’s primary concern in regards to Mr. Bryant, it certainly crossed everyone’s mind at some point. The documentation of his physician was clear and thorough, unusually so for a busy clinician. Many attempts were made to offer alternative options for care, including different antibiotic regimens and less rigorous monitoring. Appropriate consultations with a psychiatrist were employed early and often. Before he signed out AMA, a bioethics panel reviewed the case. By any reasonable standards, an unusual and commendable amount of care and effort was provided to Mr. Bryant. Yet, despite these measures, there was no happy ending to his case.

Less than two days after he was readmitted to the hospital, Mr. Bryant was found in his hospital room unresponsive. One minute he was talking with a nurse, and the next minute he was dead, likely the victim of a massive pulmonary embolism. One wonders if it makes any difference then, that he called all the shots, right up until his last breath. The day before, an ultrasound had shown several blood clots in his legs. Yet despite conversations with several doctors, he refused anticoagulation with heparin, the one treatment that might have saved him. At no point was any treatment given without his approval, and no refusals were made without acknowledging the risks and alternatives. Mr. Bryant was truly a man who stood tall, if not proud, in the face of risks and went against medical advice. Or was it to spite it? That, unfortunately, we will never know.


Anis AH, Sun H, Guh DP, Palepu A, Schechter MT, O’Shaughnessy MV. 2002. Leaving hospital against medical advice among HIV-positive patients. CMAJ 167(6): 633-7.

Brook M, Hilty DM, Liu W, Rona H, Frye MA. 2006. Discharge against medical advice from inpatient psychiatric treatment: a literature review. Psychiatr Serv. 57: 1192-8.

Devitt PJ, Devitt AC, Dewan M. 2000. Does identifying a discharge as “against medical advice” confer legal protection? J Fam Pract. 49(3): 224-7.

Huang SW, Li J, Gupta R, Chien V, Martin, RE. 2003. What happens to patients who leave hospital against medical advice? CMAJ 168(4): 417-20.

Jeremiah J, O’Sullivan P, Stein MD. 1995. Who leaves against medical advice? J Gen Intern Med. 10(7): 403-5.

Pages KP, Russo JE, Wingerson DK, Ries RK, Roy-Byrne PP, Cowley DS. 1998. Predictors and outcomes of discharge against medical advice from the psychiatric units of a general hospital. Psychatr Serv. 49(9): 1187-92.

Saitz R, Ghali WA, Moskowitz MA. 2000. The impact of leaving against medical advice on hospital resource utilization. J Gen Intern Med. 15(2): 103-7.

Weingart SN, Davis RB, Phillips RS. 1998. Patients discharged against medical advice from a general medicine service. J Gen Intern Med. 13(8): 568-71.a