Case and Commentary
Aug 2008

Informed Refusal in the Emergency Department: Is It Really Informed?

Matthew R. Lewin, MD, PhD
Virtual Mentor. 2008;10(8):487-491. doi: 10.1001/virtualmentor.2008.10.8.ccas1-0808.


Dr. Padgett was the attending physician during the night shift in an urban hospital emergency department. He had one resident and one intern working with him. Over the span of 4 hours, he had seen 17 patients. There were 21 patients still in the waiting room when an ambulance arrived carrying Ms. Burton, a 38-year-old professional woman, who had been drinking alone. As she exited the bar, bystanders witnessed her fall from standing height, hitting her forehead. The EMS run sheet said it was uncertain whether she'd had any loss of consciousness. Witnesses could only attest to the fact that, if she had lost consciousness, it was a brief loss because they heard her muttering expletives on the sidewalk seconds later. Her prehospital fingerstick glucose was normal. She arrived in the emergency department on a spine board, wearing a cervical collar, and very angry about being taken to the hospital against her will.

Ms. Burton was perfectly oriented, alert, intoxicated on alcohol, and combative with staff, refusing care. She had signs of a fall—a contusion and abrasion to her left orbit and forehead. She growled her displeasure as Dr. Padgett checked her vital signs, which were unremarkable except for a modest elevated blood pressure and tachycardia. To the extent that a neurological exam could be performed, it was nonfocal.

After the exam, when Dr. Padgett told Ms. Burton that he was ordering a CT scan of her head to check for intracranial bleeding, she refused. Dr. Padgett began explaining the reasons for his decision and the consequences of bleeding in her skull, but Ms. Burton was having none of it, claiming that she couldn’t afford those tests. She removed the IV from her arm, flung it in the direction of Dr. Padgett, and attempted to clamber over the railings of the gurney only to be restrained by nurses and staff concerned about her fall risk.

Ms. Burton admitted, by now, to having a headache at the site of the forehead contusion, but said she'd go home and come back immediately if she "got worse." She steadfastly denied depression, suicidal ideation, or even alcohol abuse, simply stating that she only drank on weekends, that she was a professional, and that this had happened "many times before." She refused to allow phone calls to family, friends, or co-workers claiming that this was "none of their business."


Let's see. I have a combative patient who may or may not have a significant brain injury and a busy, understaffed emergency department. The patient appears to have decision-making capacity, but has trauma above the clavicle, a headache despite intoxication, and is irritable and impatient—all classic signs of acute brain injury. We haven’t even considered other elements of the differential diagnosis for these signs, which include electrolyte abnormalities, arrhythmias, cerebellar dysfunction, normal pressure hydrocephalus, Wernicke’s encephalopathy, toxic alcohols, and infection. Ms. Burton may have a second diagnosis that is equally important and potentially discoverable simply because she fell.

I have several options. Option 1: Try to reason with her and then let her go at her own peril (assuming she can walk). I might ask her if she would accept the CT scan if she weren’t charged for it. If the answer is “No, let me go,” that would call her judgment into question because she is contradicting her stated reason for refusing care. In that case, I could put her on a legal, medical hold for being a danger to herself—a weak argument to an informed consent “absolutist,” for whom the patient’s reasons for refusing don’t really matter as long as she has demonstrated the ability to reason and is oriented. I could also call a psychiatrist to assess her decisionmaking capacity, but the psychiatrist will confront the same problem I have of sorting out the irritability from acute brain injury, something I am far more likely to have seen and with greater frequency than the average psychiatrist. Nevertheless, the patient might sober up even before the psychiatrist’s assessment and benefit from the consultation that would at least touch on substance abuse.

I am completely opposed to this two-part option—letting the patient check out against medical advice or seeking a psychiatric consult. We don’t have the time or resources to wait for a consultant when the patient isn’t even medically cleared for psychiatric assessment. The patient will receive referrals for psychiatric and substance abuse counseling, whether she sees a psychiatrist tonight or not. If she leaves the department against medical advice, she will still be responsible for the ambulance bill, facility and nursing charges, and the medical screening exam by Dr. Padgett. Although the hospital will be willing to accept some type of payment schedule, it will be hard for Ms. Burton to argue that she is being charged for care that was unreasonable, considering her risk factors for significant brain injury.

Option 2: Do nothing except observe her. Considering her determination to leave, this would require commitment from the staff to keep her in her gurney whether or not she was on a medical hold. Restrained patients have to be monitored at all times, and intoxicated patients can’t be walking around the emergency department since they are fall risks, something this patient has already proven. I could call the nursing supervisor to see if there are staff members able and willing to float down to the ED to help out. Alternatively, there might be security staff credentialed as sitters. If she sobers up, can walk, and still has decision-making capacity, then it should be safe for her to go home, even if she lives alone. This approach is commonly referred to as “metabolizing to freedom.” Given that most brain injuries are not, in fact, neurosurgical emergencies, even when there is blood, a respectable minority would argue that it is not even important to identify all bleeds—especially if the patient’s mentation clears.

This option is more appealing in the sense that the patient is potentially spared the expense of the test she says she doesn't want, but we still get to observe her. If her behavior improves at a rate within some reasonable range for alcohol intoxication, she can be reexamined and released. The problem is that she is actively trying to "escape," and it is not clear whether her attempts are prompted by a quite rational fear of the bill she will receive or whether she is suffering from an injured brain. Bills from the radiology department will be much more difficult to negotiate than mine and the emergency department's facility and nursing fees. I can certainly cancel my pro-fee and reduce the ED charges by advocating for the patient, but I cannot ask the radiology department to cut their charges for the test I order. The patient could pay the negotiated fees over an extended period of time, if she needs further financial assistance. I routinely write letters on behalf of self-insured patients. Why should they pay the "billed" rates that are discounted to third-party payers and government programs such as Medicaid? In my experience, the hospitals are happy to collect less in exchange for less trouble.

Option 3: Restrain, sedate and get a CT scan of Ms. Burton’s brain. This option places the interests of the emergency department and its patients as a whole over the immediate interests of the individual patient. This patient is a distraction to my staff and to me that comes at the expense of others’ care and safety. Acting on this option resolves the medical questions quickly, spares resources (e.g., staff and bed) for the benefit of others, and most safely returns the patient to her home or the street—but it completely ignores the patient’s stated wishes. If her brain CT is normal, she can refuse imaging of her cervical spine (heretofore neglected in the discussion, but definitely indicated). She can remain in a cervical collar until the dust settles. If she has a clinically significant fracture, that isn’t likely to be the reason she is agitated. If her CT scan is positive, then we would proceed with cervical spine imaging and not ask her permission.

Dr. Padgett has been "on" for at least 4 hours and probably has no more than 4 to 8 hours to resolve the case of Ms. Burton before he will have to sign her out to the next attending. Sign-out is notoriously dangerous under the best of circumstances, and Dr. Padgett will be signing out a patient with altered mental status to the next attending, Dr. M'Fleur, who did not see the patient at presentation or during any of the repeat neurological examinations. As far as Dr. M'Fleur is concerned, this is a new patient with a new baseline. If Dr. Padgett elected to use chemical or physical restraints or both, Ms. Burton's neurological examination could be compromised by sedation or she may simply be hung-over and sleepy. How is Dr. M'Fleur supposed to assess the patient or have any idea about the rate at which Ms. Burton should be improving? Dr. M'Fleur is inheriting a high-risk situation, even if the patient hadn't refused care, and Dr. Padgett was merely following the "metabolize to freedom" strategy.

As a general rule, I have a lower threshold for CT scanning if I am signing out an intoxicated patient. It is not appropriate to risk patient safety and foist diagnostic dilemmas or my clinical style onto the new team.

Dr. Padgett's medical decision making has to account for his presence or absence at the time he anticipates Ms. Burton's disposition home or admission for observation by the neurosurgical or medicine services. Supposing the ED is so busy or Dr. Padgett is so tied up with other patients that he cannot return to the bedside for reassessment of Ms. Burton, resulting in a potentially dangerous delay in her care? Post hoc critiques of his care will call his judgment into question: Wasn't it obvious that she had a bleed?

Scarce Resources

Allocation of scarce resources must also be considered. Every decision the physician makes about one patient has a consequence for every other patient in the emergency department. Every action, procedure, phone call or order consumes nursing time, clerical time, adds to the physician’s “to do” list, and, ultimately, affects patient flow through the department. Thus, practical considerations inevitably come into play with any patient—especially challenging ones like Ms. Burton.

Much has been written about the differences between emergency medicine and primary care practice. Among many factors, emergency department personnel do not usually know the patient or their families; the patient has experienced an acute change in health; decisions are made quickly and independent of outside information or consultation with others who do know the patient; the physician is working in an open and uncontrolled environment; and anxiety, pain, alcohol, and altered mental status are frequent occurrences, typically with a high frequency of serious, underlying pathology. If there is no way for the emergency physician to buy enough time to make a safe decision for the patient or avoid compromising the safety of others, the physician must make a decision which is, ideally: (1) based on impartiality, (2) universalizable to other situations, and (3) what others would likely find justifiable given the same set of circumstances. This is a sort of formalized "golden rule."

I don't know what Dr. Padgett should do, but I know what I would do, absent the luxury of time, personnel, and a family member of Ms. Burton whom I could contact. I would sedate or restrain Ms. Burton, or both, image her brain, and hope it is normal. She would be discharged with written instructions, referrals, and a personal note from me that if she cannot pay the full bill, I will advocate for significant reduction. After that, it is her choice and responsibility to take me up on my offer to write a letter on her behalf. She is free to file a complaint against me, which might help reduce her bill even further or erase it if she is sufficiently resourceful.

Given more resources or time I would observe her. Given a responsible family member or close friend who says she is "always this way," when intoxicated, and if she has not indicated that she has any intention to harm herself, I would honor her wishes without much hesitation.

Did I mention that I saw this patient last night…?


Virtual Mentor. 2008;10(8):487-491.



The people and events in this case are fictional. Resemblance to real events or to names of people, living or dead, is entirely coincidental. The viewpoints expressed on this site are those of the authors and do not necessarily reflect the views and policies of the AMA.