“I Need to Go—Today”: Reevaluating Our Approach to Against Medical Advice Discharges

My pager hummed against my waist one morning as I gathered my things to begin rounds. “Mr. Smith is requesting to leave immediately,” the text read. “Please come to the bedside.”

I entered Mr. Smith’s room for the third day in a row, eager to convince him to stay. “I need to go—today,” he said immediately. “My dog is really sick and no one can take care of him. I know, the nurse told me if I left against medical advice, I might have to sign some papers and my hospital stay might not be covered, but my dog is all I have, and I have to go—today.”

Mr. Smith was about halfway through the expected length of his hospital stay in which he was receiving intravenous antibiotics to treat bacteremia. Because of his recent intravenous drug use, he wasn’t considered a good candidate for home parenteral therapy. “I have to go,” he kept insisting. “I can come back in a few days if you want me to, but I just can’t stay today.”  

I felt for him. As a man with housing insecurity who occasionally lived out of his car with his dog, Pepper, Mr. Smith was in agony over the time he had already spent away from his closest companion. Just the day before, we shared pictures of our dogs and bonded over their likeness. With Pepper now feeling ill, Mr. Smith felt compelled to go and care for him despite his own ailment. “Okay,” I finally lamented. “But you’ll have to sign the AMA paperwork.”

“AMA.” Used as both a noun and a verb, this shorthand refers to a patient who wishes to be discharged from the hospital against medical advice (AMA). The designation is generally given to patients who request discharge before it is deemed safe or ideal by the provider. Without a standardized approach or method to AMA discharges, residents basically learn their preferred strategy of implementation from their attendings and peers. Somewhere along my own training, I was taught that if a patient had the capacity to make a decision (with little formal training on how to evaluate such capacity), then they could go. We were tasked with advising them of all the risks (which somehow always included death) and having them sign a dedicated AMA form, which implied protection from legal retribution for subsequent poor outcomes. I was falsely told by residents and attendings alike that if patients left AMA, insurance would not pay for their hospitalization. Both the threat of death and significant financial burden were routinely used to coerce patients into following our plan, without sharing evidence to back up those claims.  

There was little emphasis in my formal training on pulling up a chair and assessing a patient’s needs—and often, their fears. Why is Pepper so important to Mr. Smith such that he would want to leave the hospital before it was safe to do so? Some may say Pepper’s just a dog, but for a patient with minimal social support, financial and housing instability, and substance use, this dog was clearly his most constant and grounding presence. It may not be what I would have done if I were Mr. Smith, but I am not him.

In his Annals of Internal Medicine article, Dr. Alfandre (1) challenges us to reevaluate our preconceived notions around the AMA process and reminds us that the benefits are unproven and that bias and harm may be present. In particular, he notes that the use of AMA terminology offers no proven benefit of liability protection; rather, it is the thorough and accurate documentation of the discharge process—which is already expected of us—that may offer some cover. Furthermore, attempting to formalize an AMA discharge by requesting signatures on a standardized document has not been proven to offer liability protection, nor improve the discussion around the reasons for discharge. More important, using such tactics does not absolve a physician from providing and executing an appropriate treatment plan at the time of discharge and instead could antagonize what may already be a fragile therapeutic relationship.

It warrants mentioning that AMA discharges are more often associated with African Americans, persons experiencing homelessness, patients with mental illness, patients with substance use, and those with Medicaid or no insurance (2). Anecdotally, this correlates with my own experience with AMA discharge. Although a variety of factors may be involved with such discrepancies, I wonder if both the systemic and implicit biases against these populations play a role in the care that leads up to and ultimately culminates with an AMA discharge. It especially concerns me that the aftermath of AMA discharge, including high rates of readmission and less aggressive treatment plans (3), may further perpetuate the health care discrepancies these populations face.  

If I could redo my interaction with Mr. Smith that day, there is a lot I would change, and a lot I have changed in subsequent similar encounters. Running into a room to have a lengthy conversation about a premature discharge is not the thing I look forward to the most, and certainly not the thing I do the best. However, over time, I have recognized that most patients who request discharge early do so for a variety of reasons that have little do with me personally. Whether it be fear or anxiety, concern about finances and housing, care of dependents, or addiction, patients have needs that they feel at times may trump their health. In counseling these patients on the benefits of continued admission, the primary goal is not liability protection but rather harm reduction. True shared decision making relies on a strictly patient-centered approach. It does not benefit from an exasperated attitude of personal offense when a patient doesn’t do what I want them to. Furthermore, taking stock of my own biases that may trigger such feelings of personal frustration may be helpful in approaching these scenarios with more empathy and understanding.

I met Mr. Smith again a few weeks later after he was readmitted to the hospital for recurrent fever. His dog was doing much better, he said. He showed me pictures for a few minutes before we switched gears to address his health. “I’m ready to stay this time,” he said.

References 

1. Alfandre D. Annals for Hospitalists inpatient notes—challenging the myths of the against medical advice discharge. Ann Intern Med. 2021;174:HO2-HO3. [PMID: 34662167] doi:10.7326/M21-3450
2. Alfandre D, Brenner J, Onukwugha E. Against medical advice discharges. J Hosp Med. 2017;12:843-845. [PMID: 28991952] doi:10.12788/jhm.27962.
3. Alfandre DJ. "I'm going home": discharges against medical advice. Mayo Clin Proc. 2009;84:255-60. [PMID: 19252113] doi:10.1016/S0025-6196(11)61143-9